International Journal of Molecular Sciences

Article Preparation of Peptide and Recombinant Tissue Plasminogen Activator Conjugated Poly(Lactic-Co-Glycolic Acid) (PLGA) Magnetic Nanoparticles for Dual Targeted Thrombolytic Therapy

Huai-An Chen 1, Yunn-Hwa Ma 2 , Tzu-Yuan Hsu 2 and Jyh-Ping Chen 1,3,4,5,*

1 Department of Chemical and Materials and Materials Engineering, Chang Gung University, Kwei-San, Taoyuan 33302, Taiwan; [email protected] 2 Department of Physiology and Pharmacology and Healthy Aging Research Center, Chang Gung University, Kwei-San, Taoyuan 33302, Taiwan; [email protected] (Y.-H.M.); [email protected] (T.-Y.H.) 3 Department of Plastic and Reconstructive Surgery and Craniofacial Research Center, Chang Gung Memorial Hospital, Linkou, Kwei-San, Taoyuan 33305, Taiwan 4 Research Center for Food and Cosmetic Safety, Research Center for Chinese Herbal Medicine, College of Human Ecology, Chang Gung University of Science and Technology, Taoyuan 33302, Taiwan 5 Department of Materials Engineering, Ming Chi University of Technology, Tai-Shan, New Taipei City 24301, Taiwan * Correspondence: [email protected]; Tel.: +886-3211-8800 (ext. 5298)



Received: 10 March 2020; Accepted: 10 April 2020; Published: 13 April 2020


Abstract: Recombinant tissue plasminogen activator (rtPA) is the only thrombolytic agent that has been approved by the FDA for treatment of ischemic stroke. However, a high dose intravenous infusion is required to maintain effective drug concentration, owing to the short half-life of the thrombolytic drug, whereas a momentous limitation is the risk of bleeding. We envision a dual targeted strategy for rtPA delivery will be feasible to minimize the required dose of rtPA for treatment. For this purpose, rtPA and fibrin-avid peptide were co-immobilized to poly(lactic-co-glycolic acid) (PLGA) magnetic nanoparticles (PMNP) to prepare peptide/rtPA conjugated PMNPs (pPMNP-rtPA). During preparation, PMNP was first surface modified with avidin, which could interact with biotin. This is followed by binding PMNP-avidin with biotin-PEG-rtPA (or biotin-PEG-peptide), which was prepared beforehand by binding rtPA (or peptide) to biotin-PEG-maleimide while using click chemistry between maleimide and the single –SH group in rtPA (or peptide). The physicochemical property characterization indicated the successful preparation of the magnetic nanoparticles with full retention of rtPA fibrinolysis activity, while biological response studies underlined the high biocompatibility of all magnetic nanoparticles from cytotoxicity and hemolysis assays in vitro. The magnetic guidance and fibrin binding effects were also confirmed, which led to a higher thrombolysis rate in vitro using PMNP-rtPA or pPMNP-rtPA when compared to free rtPA after static or dynamic incubation with blood clots. Using pressure-dependent clot lysis model in a flow system, dual targeted pPMNP-rtPA could reduce the clot lysis time for reperfusion by 40% when compared to free rtPA at the same drug dosage. From in vivo targeted thrombolysis in a rat embolic model, pPMNP-rtPA was used at 20% of free rtPA dosage to restore the iliac blood flow in vascular thrombus that was created by injecting a blood clot to the hind limb area.

Keywords: magnetic nanoparticles; tissue plasminogen activator; poly(lactic-co-glycolic acid); targeted drug delivery; clot lysis; nanomedicine

Int. J. Mol. Sci. 2020, 21, 2690; doi:10.3390/ijms21082690 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 2690 2 of 23

1. Introduction Within the two ischemic and hemorrhagic categories of stroke, the ischemic stroke accounts for ~87% of all cases, according to the American Stroke Association [1]. A multitude of molecular pathways may be involved in the onset and progression of ischemic stroke, thus an equally diverse arsenal of intervention strategies is needed. To date, the gold standard for intravenous (IV) intervention of ischemic stroke is by administrating recombinant tissue plasminogen activator (rtPA), a thrombolytic drug that dissolves clots to restore blood flow. As a serine protease, rtPA is the major enzyme that is responsible for clot dissolution by catalyzing the conversion of plasminogen to plasmin. It is currently the only thrombolytic agent approved by the U.S. Food and Drug Administration (FDA) for the treatment of ischemic stroke [2]. A major issue in rtPA therapeutic use is the short half-life (4–8 min) in the circulation due to autolysis and the influences of inhibitors, enzymes, and antibodies in blood [3]. The poor affinity of rtPA toward thrombus also reduces its thrombolysis efficacy and might cause ischemia reperfusion injury to the neuron beyond the thrombolysis window [4]. Other limitations of rtPA for clinical use include the short window time for treatment (usually within 3 h) and the risk of hemorrhagic side effects [5]. Consequently, several strategies have emerged to improve the thrombolytic efficacy of rtPA, such as ultrasound-based thrombolysis, targeting thrombolysis, and immobilized thrombolytic drug to improve the safety and effectiveness of thrombolytic therapy [6]. Nanomedicine using nanoparticles for drug delivery is a novel field for the diagnosis and treatment of diseases. Although with similar biologic molecular scale when compared with traditional medicine, the unique properties of nanoparticles provide more strategic advantages and application flexibility over pure molecular therapeutics [7]. Using nanoparticles as rtPA carrier for thrombolytic therapy has also been widely explored. For example, biodegradable polymers conjugated with rtPA can provide the protection form the inhibitor in the circulation to prolong its circulation time, which might also confer the possibility for controlled release in encapsulated drug formulation [8]. In recent years, magnetic nanoparticles (MNP), especially iron oxide (Fe3O4) MNP, are frequently employed as a nano-carrier for drug delivery. Not only to be useful as a magnetic resonance imaging contrast agent, the MNP can also be endowed with other advantages, such as magnetic targeting (physical targeting) and localized heating by magnetic field induction or near-infrared laser irradiation [9]. In addition to pristine nanoparticles, MNP could be also surface modified to enhance their functionality as a drug delivery vehicle. The most common example is surface modification with polyethylene glycol (PEG) to prevent uptake by the mononuclear phagocyte system during circulation, thus prolonging the life time of injected drug [10]. Conjugation with ligand is another commonly used strategy for nanoparticle-based targeted drug delivery. Indeed, ligand-mediated drug delivery could offer a highly specific binding between the ligand molecule immobilized to the nano-carrier and the receptor molecule highly expressed in the diseased area, which could enhance the treatment efficacy by inducing drug accumulation and increasing local drug concentration in the targeted area [11]. Many groups have proposed nanomedicine using targeted rtPA delivery strategy for thrombolysis [12]. We have extensively studied different MNP-based polymeric nano-carrier for magnetically targeted delivery of rtPA, being prepared either by immobilization of rtPA on the particle surface or by encapsulation within the polymeric matrix, and demonstrated the improved thrombolytic efficacy both in vitro and in vivo [13–15]. Recently, we also pioneered the use of thermosensitive magnetic liposomes for targeted delivery and temperature-sensitive release of rtPA [16,17]. Other groups used different ligands for targeting fibrin in a blood clot or different moieties that are associated with the blood clot. For example, the ligands used for rtPA delivery targeting thrombus include peptide targeting FXIII [18], fibrin antibody targeting fibrin [19], Arg-Gly-Asp (RGD) [20], cyclicRGD [21], or other peptides [22] targeting GP IIb/IIIa, and fucoidan targeting p-selectin [23]. Those targeted thrombolysis strategies have demonstrated promising results in improving thrombolytic efficiency with shorter clot lysis time, which makes possible the use of reduced drug dosage to circumvent associated side effects, such as the bleeding risk [24]. Int. J. Mol. Sci. 2020, 21, 2690 3 of 23

Consider a ligand targeting fibrin, which is a trimeric molecule consisting of α-, β-, and γ-chains and a major constituent of fresh or old blood clots. Although the actual quantity of fibrin content in a clot varies from clot to clot, fibrin is the major component in a clot and it exists on the surface of a clot that is slowly dissolving either spontaneously or during therapeutic thrombolytic intervention. It is conceivable that a peptide specific for fibrin could be labelled with radioisotopes and used as an imaging agent to detect vascular thrombosis. This was demonstrated from a previous study using 99mTc labelled pentapeptide Gly-Pro-Arg-Pro-Pro (GPRPP), which has high affinity for the fibrin α-chain, for the imaging of vascular thrombosis in animals [25]. The peptide GGSKGC was later added to the C-terminus of GPRPP in order to impart functionality on the peptide. This peptide (GPRPPGGSKGC) was later shown to have high fibrin-avid affinity and high resistance to proteolysis [26]. Although using a magnetic field to magnetically guide the movement of MNP for rtPA delivery is promising, the optimization of a magnetic field might be the limiting factor, especially for clinical application. On the other hand, ligand targeting for rtPA delivery might pose challenges, such as specificity and possible immunogenicity of the targeting moiety. Therefore, a dual targeting mechanism for rtPA and for other thrombolytic drugs delivery in general is warranted. This dual targeted strategy for rtPA delivery has rarely been investigated [21]. A recent paper using two peptides as ligands to construct dual targeted liposomes for delivery of a neuroprotective agent proved to be a safe and effective treatment for ischemic stroke [27]. By integrating different targeting functionalities, dual and even multi-targeted nanoparticles could be designed to provide a paradigm for precise drug delivery to the targeted site. Through sequential or synchronized navigation of the drug-loaded nano-sized carrier, highly controllable drug delivery by synergistic function of the targeting moieties could be accomplished when compared to conventional targeting strategies [28]. Therefore, with the aim of enhancing thrombolysis efficacy, we postulated that a dual targeted rtPA nano-drug would be feasible to achieve better thrombolytic outcomes for translation into a clinical setting. When compared with the single targeting strategy, nanoparticles engineered with dual targeting mechanism, either augment the single targeting ligand with a second ligand or with magnetic targeting, has successfully been used for cancer therapy, but it has never have been demonstrated for targeted thrombolysis in vivo. Because dual-targeting nanoparticles were shown to increase differentiation between cancer and normal cells and lead to enhanced accumulation of the drug in tumors [29], the same principle is deemed to be feasible for more selectively targeted delivery of thrombolytic agents to thrombus. Specifically, magnetic PLGA nanoparticles have been explored in a dual targeted delivery of paclitaxel and curcumin in brain tumor therapy [30] and protein antigen delivery for immune stimulation [31]. In the present study, PLGA MNP (PMNP) was prepared to encapsulate Fe3O4 nanoparticles in a single oil-in-water emulsion step. Thrombolytic agent (rtPA) and fibrin-affinity peptide (GPRPPGGSKGC) were conjugated to PMNP surface that was modified with avidin to prepare pPMNP-rtPA, through avidin-biotin interaction and thiol–maleimide “click” reactions using biotin-PEG-maleimide (Figure1). The physico-chemical properties of nanoparticles during preparation were analyzed, followed by biocompatibility and thrombolysis analysis in vitro. Finally, the clot thrombolysis efficiency of pPMNP-rtPA was tested in vivo in a rat embolic model.

2. Results and Discussion

2.1. Preparation of Peptide and rtPA Conjugated PLGA Magnetic Nanoparticles Iron oxide MNP are prone to agglomeration in biological fluids due to high surface energy and the tendency to adsorb proteins, which might limit their in vivo application. After entrapping in a polymer matrix, the MNP improve stability in biological fluids and increase the half-life during circulation in vivo [32]. Therefore, in this study, PLGA was used as a polymer matrix to encapsulate hydrophobic oleic acid-coated iron oxide MNP (OMNP) and form PLGA magnetic nanoparticles (PMNP) in a single oil-in-water (O/W) emulsion step in order to prevent aggregation of iron oxide MNP. After modifying Int. J. Mol. Sci. 2020, 21, 2690 4 of 23

PLGA surface with avidin through carbodiimide-mediated covalent bond formation between the carboxyl groups of PLGA and the amine groups of avidin, avidin-biotin interaction was explored for facile synthesis of rtPA and/or peptide conjugated nanoparticles (Figure1). Avidin is a protein showing considerable affinity for biotin, which is an important co-factor in many biological processes. Avidin has the ability to bind up to four biotin molecules with the strongest non-covalent interaction being 15 known to exist between a protein and its ligand (Kd = 10− M) [33]. In addition, the bond formation between biotin and avidin is very rapid and the formed bond is unaffected by pH, organic solvents, or temperature, which facilitates the conjugation of rtPA (or peptide) to PMNP-avidin. The one-to-four binding ratio of avidin toward biotin also allows high loading of rtPA on nanoparticle surface in a controllableInt. J. Mol. Sci. manner2020, 21, 2690 to exert high thrombolytic activity. 4 of 24

FigureFigure 1. 1.The The schematic schematic diagram diagram showing showing the preparationthe preparation of peptide of peptide conjugated conjugated poly(lactic-co-glycolic poly(lactic-co- acid)glycolic (PLGA) acid) magnetic(PLGA) magnetic nanoparticles nanoparticles (pPMNP), (pPMNP), recombinant recombinant tissue plasminogen tissue plasminogen activator activator (rtPA) conjugated(rtPA) conjugated PLGA magnetic PLGA magnetic nanoparticles nanoparticles (PMNP-rtPA), (PMNP-rtPA), and peptide and/rtPA peptide/rtPA conjugated conjugated PLGA magnetic PLGA nanoparticlesmagnetic nanoparticles (pPMNP-rtPA). (pPMNP-rtPA).

2. ResultsTherefore, and weDiscussion used biotin as a mediator in this study to introduce rtPA and peptide to PMNP-avidin. This was accomplished using biotin-PEG-maleimide, which could interact with the nanoparticle through avidin2.1. Preparation/biotin interaction of Peptide in and one rtPA end Co andnjugated conjugate PLGA rtPA Magnetic (or peptide) Nanoparticles at the other using thiol–maleimide “click” chemistry [34]. It is well known that maleimide specifically reacts with the sulfhydryl group betweenIron pH oxide 6.5 toMNP 7.5 andare formsprone ato stable agglomeration irreversible inthioether biological linkage. fluids due It should to high be surface noted, however,energy and a morethe tendency alkaline conditionsto adsorb proteins, (pH > 8.5) which favors might reaction limit with their primary in vivo aminesapplication. and theAfter rate entrapping of maleimide in a hydrolysispolymer matrix, increases the [35 MNP]. The improve facile conjugation stability in rtPA biological and peptide fluids to and PMNP-avidin increase the is through half-life binding during ofcirculation rtPA (or peptide) in vivo [32]. to biotin-PEG-maleimide Therefore, in this study, with PLGA a 1:1 was molar used ratio as ofa polymer maleimide: matrix rtPA to (or encapsulate peptide). Thishydrophobic could be accomplished,oleic acid-coated as either iron GPRPPGGSKGCoxide MNP (OMNP) or rtPA and has form only PLGA one thio magnetic (–SH) group nanoparticles from the cysteine(PMNP) (C) in a residue single atoil-in-water one end of (O/W) the peptide emulsion or anstep unpaired in order cysteine to prevent residue aggregation in rtPA of (contains iron oxide 35 cysteineMNP. After residues modifying and 17 disulfidePLGA surface linkages) with [36 avidin]. through carbodiimide-mediated covalent bond formationIt is imperative between tothe first carboxyl modify groups PMNP of surface PLGA with and vast the amountamine groups of avidin of foravidin, conjugating avidin-biotin with abundantinteraction biotin-PEG-rtPA was explored for in facile order synthesis to optimize of rtPA the preparationand/or peptide of pPMNP-rtPA.conjugated nanoparticles This high specific (Figure fibrinolytic1). Avidin activityis a protein of immobilized showing considerable rtPA per unit affini weightty offor nanoparticle biotin, which will is render an important the use of co-factor minimum in many biological processes. Avidin has the ability to bind up to four biotin molecules with the strongest non-covalent interaction being known to exist between a protein and its ligand (Kd = 10−15 M) [33]. In addition, the bond formation between biotin and avidin is very rapid and the formed bond is unaffected by pH, organic solvents, or temperature, which facilitates the conjugation of rtPA (or peptide) to PMNP-avidin. The one-to-four binding ratio of avidin toward biotin also allows high loading of rtPA on nanoparticle surface in a controllable manner to exert high thrombolytic activity. Therefore, we used biotin as a mediator in this study to introduce rtPA and peptide to PMNP- avidin. This was accomplished using biotin-PEG-maleimide, which could interact with the nanoparticle through avidin/biotin interaction in one end and conjugate rtPA (or peptide) at the other using thiol–maleimide “click” chemistry [34]. It is well known that maleimide specifically reacts with the sulfhydryl group between pH 6.5 to 7.5 and forms a stable irreversible thioether linkage. It should be noted, however, a more alkaline conditions (pH > 8.5) favors reaction with primary amines and the rate of maleimide hydrolysis increases [35]. The facile conjugation rtPA and peptide to PMNP-

Int. J. Mol. Sci. 2020, 21, 2690 5 of 24

avidin is through binding of rtPA (or peptide) to biotin-PEG-maleimide with a 1:1 molar ratio of maleimide: rtPA (or peptide). This could be accomplished, as either GPRPPGGSKGC or rtPA has only one thio (–SH) group from the cysteine (C) residue at one end of the peptide or an unpaired cysteine residue in rtPA (contains 35 cysteine residues and 17 disulfide linkages) [36]. It is imperative to first modify PMNP surface with vast amount of avidin for conjugating with Int. J. Mol. Sci. 2020, 21, 2690 5 of 23 abundant biotin-PEG-rtPA in order to optimize the preparation of pPMNP-rtPA. This high specific fibrinolytic activity of immobilized rtPA per unit weight of nanoparticle will render the use of amountminimum of nanoparticles amount of nanoparticles for effective for thrombolysis effective thrombin vivoolysis. For in thisvivo. purpose, For this thepurpose, influence the influence of the amountof the ofamount avidin of used avidin for used immobilization for immobilization on the amount on the ofamount immobilized of immobilized avidin protein avidin per protein mg of per nanoparticlemg of nanoparticle was studied. was studied. The immobilized The immobilized avidin content avidin increasedcontent increased with the with amount the amount of avidin of used avidin duringused theduring modification the modification step and step reached and reached a plateau a plateau at around at around 0.2 mg 0.2 avidin, mg avidin, as shown as shown in Figure in Figure2a. Therefore,2a. Therefore, the optimum the optimum avidin avidin modification modification condition cond is reachedition is reached when 0.2 when mg avidin 0.2 mg is avidin reacted is with reacted 1 mgwith PMNP 1 mg for PMNP preparation for preparation of PMNP-avidin of PMNP-avidin where 63.5 where1.5 µ63.5g avidin ± 1.5 /μmgg avidin/mg PMNP loading PMNP content loading ± couldcontent be achieved. could be Usingachieved. this Using PMNP-avidin this PMNP-avidin preparation, preparation, we proceeded we proceeded to study eff toect study of the effect amount of the ofamount rtPA used of rtPA for immobilization used for immobilization on the amidolytic on the amidolytic activity of activity immobilized of immobilized rtPA activity rtPA per activity mg of per nanoparticle.mg of nanoparticle. The loading The activity loading increases activity increases with rtPA with used rtPA for conjugation used for conjugation until reaching until 0.5 reaching mg rtPA 0.5 andmg decreases rtPA and thereafter, decreases as thereafter, shown in Figure as shown2b. Therefore, in Figure the2b. optimum Therefore, drug the loading optimum is reached drug loading when is 0.5reached mg rtPA when is reacted 0.5 mg with rtPA 1 mgis reacted PMNP-avidin with 1 mg to fabricate PMNP-avidin PMNP-rtPA to fabricate with 0.35PMNP-rtPA0.02 U with rtPA /0.35mg ± ± PMNP-avidin0.02 U rtPA/mg loading PMNP-avidin activity. Forloading the properactivity. preparation For the proper of pPMNP-rtPA preparation of without pPMNP-rtPA overloading without PMNP-avidin,overloading PMNP-avidinPMNP-avidin, wasPMNP-avidin first reacted was with first biotin-PEG-peptide reacted with biotin-PEG-peptide corresponding tocorresponding 5 mol% of rtPAto 5 in mol% the first of step,rtPA followedin the first by step, reacting followed with by biotin-PEG-rtPA reacting with biotin-PEG-rtPA to fabricate pPMNP-rtPA to fabricate (Figure pPMNP-1). ThertPA final (Figure loading 1). ofThe peptide final loading on pPMNP-rtPA of peptide was on pPMNP-rtPA 0.46 0.03 mg was/mg 0.46 nanoparticle, ± 0.03 mg/mg while nanoparticle, the rtPA ± activitywhile slightlythe rtPA reduced activity to slightly 0.32 reduced0.01 U rtPA to 0.32/mg ± nanoparticle. 0.01 U rtPA/mg nanoparticle. ±

Figure 2. The effect of avidin and rtPA used during for immobilization on the loading of avidin protein Figure 2. The effect of avidin and rtPA used during for immobilization on the loading of avidin (a) and rtPA activity (b). protein (a) and rtPA activity (b). The activity of rtPA determined above is the amidolytic activity (U) to hydrolyze a low molecular The activity of rtPA determined above is the amidolytic activity (U) to hydrolyze a low weight substrate (S-2288), but not associated with the clot lysis thrombolytic activity. The fibrinolytic molecular weight substrate (S-2288), but not associated with the clot lysis thrombolytic activity. The property of free and immobilized rtPA was compared from the agar plate assay to show the fibrinolytic property of free and immobilized rtPA was compared from the agar plate assay to show thrombin-mediated fibrinolytic activity of rtPA for application in thrombolysis in order to ascertain the the thrombin-mediated fibrinolytic activity of rtPA for application in thrombolysis in order to retention of biological activity of rtPA, which is likely to change and diminish the clot lysis activity ascertain the retention of biological activity of rtPA, which is likely to change and diminish the clot due to protein denaturation [37]. To validate this assay method, the dosage dependent sensitivity was lysis activity due to protein denaturation [37]. To validate this assay method, the dosage dependent first confirmed by determining the thickness of clear zones around sample holes, containing different sensitivity was first confirmed by determining the thickness of clear zones around sample holes, amount of rtPA, due to the enzymatic action of the drug. The thickness of a clear zone around the containing different amount of rtPA, due to the enzymatic action of the drug. The thickness of a clear sample hole of the solidified agar, due to lysis of fibrin gel with rtPA enzymatic function, increased with zone around the sample hole of the solidified agar, due to lysis of fibrin gel with rtPA enzymatic the amount of rtPA used for the assay, as shown in Figure3a,b. When PMNP-rtPA or pPMNP-rtPA function, increased with the amount of rtPA used for the assay, as shown in Figure 3a and b. When was placed in the sample hole on the agar plate, the size of the clear zone created by the action of PMNP-rtPA or pPMNP-rtPA was placed in the sample hole on the agar plate, the size of the clear immobilized rtPA was equal to that created by free rtPA at the same drug dosage (Figure3a,c), endorsing zone created by the action of immobilized rtPA was equal to that created by free rtPA at the same that rtPA conjugated to PMNP and pPMNP could fully retain its lytic action against fibrin clot. Most drug dosage (Figure 3a and c), endorsing that rtPA conjugated to PMNP and pPMNP could fully importantly, nano-carrier conjugated with the peptide ligand (pPMNP-rtPA) showed the same clot lysis activity when compared to the non-conjugated one (PMNP-rtPA), implying that modifying PMNP with the targeting peptide ligand would not influence the thrombolysis activity of rtPA. The absence of any clear zone around the phosphate buffered saline (PBS), PMNP and pPMNP-containing sample Int. J. Mol. Sci. 2020, 21, 2690 6 of 24 retain its lytic action against fibrin clot. Most importantly, nano-carrier conjugated with the peptide ligand (pPMNP-rtPA) showed the same clot lysis activity when compared to the non-conjugated one (PMNP-rtPA), implying that modifying PMNP with the targeting peptide ligand would not influence theInt. J.thrombolysis Mol. Sci. 2020, 21 activity, 2690 of rtPA. The absence of any clear zone around the phosphate buffered saline6 of 23 (PBS), PMNP and pPMNP-containing sample hole shows that they have no fibrinolytic activity to hydrolyzehole shows the that fibrin they and have the no lysis fibrinolytic of agar plate activity is so tolely hydrolyze due to the the fibrinolytic fibrin and theactivity lysis of of rtPA agar (Figure plate is 3a).solely due to the fibrinolytic activity of rtPA (Figure3a).

Figure 3. The clot lysis ability of free and immobilized rtPA form the fibrinolytic activity assay using the Figure 3. The clot lysis ability of free and immobilized rtPA form the fibrinolytic activity assay using agar plate. Fibrinolysis activity was observed from a clear zone around the sample hole after incubation the agar plate. Fibrinolysis activity was observed from a clear zone around the sample hole after at 37 C for 24 h using free rtPA at 10, 20, and 30 µU and PMNP-rtPA and pPMNP-rtPA at 20 µU(a). incubation◦ at 37 °C for 24 h using free rtPA at 10, 20, and 30 μU and PMNP-rtPA and pPMNP-rtPA at The thickness of the clear zone (blue lines) in (a) was measured and compared between free rtPA at 20 μU (a). The thickness of the clear zone (blue lines) in (a) was measured and compared between free different dosage (b), and between free rtPA, PMNP-rtPA, and pPMNP-rtPA (c). (d) Comparison of the rtPA at different dosage (b), and between free rtPA, PMNP-rtPA, and pPMNP-rtPA (c). (d) fibrinolytic activity of phosphate buffered saline (PBS) (control), rtPA, PMNP-rtPA and pPMNP-rtPA Comparison of the fibrinolytic activity of phosphate buffered saline (PBS) (control), rtPA, PMNP-rtPA from the lysis index using thromboelastometry at 1.7 µU/mL rtPA dosage. (e) The inhibition of free and pPMNP-rtPA from the lysis index using thromboelastometry at 1.7 μU/mL rtPA dosage. (e) The rtPA, PMNP-rtPA, and pPMNP-rtPA by plasminogen activator inhibitor-1 (PAI-1) in vitro by measuring inhibition of free rtPA, PMNP-rtPA, and pPMNP-rtPA by plasminogen activator inhibitor-1 (PAI-1) the residual amidolytic activity of rtPA after incubating with PAI-I at 37 ◦C for different times. in vitro by measuring the residual amidolytic activity of rtPA after incubating with PAI-I at 37 °C for differentWe further times. used thromboelastometry to determine blood clots lysis induced by rtPA by measuring the lysis index, which was shown to be dosage-dependent on free rtPA concentration used for the assay, to evaluate the thrombolytic effects [38]. As shown in Figure3d, there is no significant di fference in lysis index among free rtPA, PMN-rtPA and pPMNP rtPA irrespective of assay time (30, 45, 60 and 120 min) at the same rtPA drug dosage, which is in accordance with those from agar plate assays Int. J. Mol. Sci. 2020, 21, 2690 7 of 24

We further used thromboelastometry to determine blood clots lysis induced by rtPA by measuring the lysis index, which was shown to be dosage-dependent on free rtPA concentration used for the assay, to evaluate the thrombolytic effects [38]. As shown in Figure 3d, there is no significant difference in lysis index among free rtPA , PMN-rtPA and pPMNP rtPA irrespective of assay time (30, 45, 60 and 120 min.) at the same rtPA drug dosage, which is in accordance with those from agar Int. J. Mol. Sci. 2020, 21, 2690 7 of 23 plate assays (Figure 3c). Taken together, both assays confirm the full retention of rtPA fibrinolytic activity after conjugation to PMNP, irrespective of peptide conjugation by using the pre-designed (Figureconjugation3c). Takenstep as together, shown in both Figure assays 1, which confirm facili thetate full the retention use of pPMNP-rtPA of rtPA fibrinolytic for dual targeted activity afterrtPA conjugationdelivery. to PMNP, irrespective of peptide conjugation by using the pre-designed conjugation step as shownThe plasminogen in Figure1, which activator facilitate inhibitor-1 the use of(PAI-1) pPMNP-rtPA is the main for dual inhibitor targeted of rtPAcirculating delivery. rtPA and inactivationThe plasminogen of rtPA by PAI-1 activator is the inhibitor-1 main cause (PAI-1) for theis short the half-life main inhibitor of rtPA [39]. of circulating Therefore, rtPA it will and be inactivationhighly desirable of rtPA that bythe PAI-1 PMNP-rtPA is the main and causepPMNP-rtPA for the short developed half-life in ofthis rtPA study [39 ].will Therefore, not only itretain will beits fibrinolytic highly desirable activity that (Figure the PMNP-rtPA 3c), but also and prolong pPMNP-rtPA the circulation developed time of in rtPA. this studyTherefore, will notwe used only retainthe chromogenic its fibrinolytic substrate activity S-2288 (Figure to determine3c), but also the prolong amidolytic the activity circulation of PMNP-rtPA time of rtPA. and Therefore, pPMNP- wertPA used when the pre-incubated chromogenic substratewith PAI-1 S-2288 for different to determine times theand amidolytic compared activitywith that of of PMNP-rtPA free rtPA. andThe pPMNP-rtPAimmobilized rtPA when was pre-incubated inactivatedwith by PAI-1 PAI-1 more for di slowlyfferent than times free and rtPA, compared indicating with thatthat ofthe free rtPA rtPA. in ThePMNP-rtPA immobilized and pPMNP-rtPA rtPA was inactivated was protected by PAI-1 from more PAI-1 slowly to thansome free extent, rtPA, as indicating shown in thatFigure the 3e. rtPA This in PMNP-rtPAmight be due and to pPMNP-rtPAthe effects of steric was protected hindrance from of the PAI-1 nano-carrier to some extent, for accessi as shownbility in of Figure PAI-13 e.to Thisbind mightrtPA or be due due to to the the inaccessibility effects of steric of hindrance the PAI-1-binding of the nano-carrier site in the for bound accessibility rtPA [40]. of PAI-1 to bind rtPA or due to the inaccessibility of the PAI-1-binding site in the bound rtPA [40]. 2.2. Characterization of Physico-Chemical Properties 2.2. Characterization of Physico-Chemical Properties From the transmission electron microscope (TEM) images that are shown in Figure 4a to d, iron oxideFrom MNP the and transmission OMNP agglomerated electron microscope in water at (TEM) low ma imagesgnification that aredue shown to Van in der Figure Waals4a–d, force iron or oxideweak MNPmagnetic and OMNPattractions agglomerated from magnetic in water dipole–d at lowipole magnification interactions due [41]. to Van Nonetheless, der Waals force the average or weak magneticparticle size attractions of discrete from nanoparticles magnetic dipole–dipole could still interactions be estimated [41]. to Nonetheless, be around the10 averagenm from particle high sizemagnification of discrete micrographs nanoparticles and could consistent still be estimatedwith similar to beresearches around 10that nm also from used high co-precipitation magnification micrographsmethod to prepare and consistent iron oxide with MNP similar [42]. researches After the encapsulation that also used co-precipitationof OMNP in PLGA, method PMNP to prepare shows ironsmooth oxide and MNP spherical [42]. Aftersurface the morpholo encapsulationgy of ~300 of OMNP nm diameter in PLGA, and PMNP an agglomerate shows smooth core and of iron spherical oxide surfaceMNPs (Figure morphology 4e and of f). ~300 This nm indicates diameter that and all an OMNP agglomerate were corewell ofencapsulated iron oxide MNPs within (Figure the PLGA4e,f). Thismatrix indicates after a thatsingle all oil-in-water OMNP were (o/w) well encapsulatedemulsion/solvent within evaporation the PLGA matrixstep. Modification after a single with oil-in-water avidin (o(PMNP-avidin),/w) emulsion/solvent conjugation evaporation with rtPA step. (PMNP-rtPA), Modification and with conjugation avidin (PMNP-avidin), with peptide conjugation (pPMNP-rtPA) with rtPAdo not (PMNP-rtPA), lead to much and change conjugation of particle with size peptide and (pPMNP-rtPA)surface morphology do not (Figure lead to much4g to l). change Nonetheless, of particle a sizegray and layer surface around morphology the round (Figure contours4g–l). from Nonetheless, 1% phosphotungstic a gray layer acid around staining, the round which contours binds to from the 1%basic phosphotungstic groups (lysine acidand staining,arginine whichresidues) binds of toproteins, the basic suggests groups that (lysine both and avidin arginine and residues) rtPA were of proteins,successfully suggests immobilized that both on avidin the nanoparticle and rtPA were surface. successfully immobilized on the nanoparticle surface.

Figure 4. TheThe transmission transmission electron microscope (TEM) images of Fe3O4 magneticmagnetic nanoparticles nanoparticles (MNP) (MNP) (a,,b), oleicoleic acid coated magnetic nanoparticles (OMNP) (c,d), PLGA magnetic nanoparticles (PMNP) (e,,f), avidin-modifiedavidin-modified PLGA magnetic nanoparticles (PMNP-avidin) ( g,h), rtPA-conjugated PLGA magnetic nanoparticlesnanoparticles (PMNP-rtPA) (PMNP-rtPA) (i,j), and(i,j), peptide and /rtPA-conjugatedpeptide/rtPA-conjugated PLGA magnetic PLGA nanoparticles magnetic (pPMNP-rtPA)nanoparticles (pPMNP-rtPA) (k,l). Bar = 2 µ (mk,l ().e, gBar,i,k =), 2 1 μµmm ( (ea,g,c,)i, andk), 1 100 μm nm(a,c ()b and,d,f ,100h,j,l nm). (b,d,f,h,j,l).

From the dynamic light scattering (DLS) results that are shown in Figure5a, the average hydrodynamic diameters for MNP, OMNP, PMNP, PMNP-avidin, PMNP-rtPA, and pPMNP-rtPA are 230.7, 229.4, 252.7, 278.8, 291.2, and 321.1 nm, respectively (Table1) and consistent with the trend observed from TEM (Figure4). The particle size constantly increased during the sequential Int. J. Mol. Sci. 2020, 21, 2690 8 of 24

From the dynamic light scattering (DLS) results that are shown in Figure 5a, the average hydrodynamic diameters for MNP, OMNP, PMNP, PMNP-avidin, PMNP-rtPA, and pPMNP-rtPA Int. J. Mol. Sci. 2020, 21, 2690 8 of 23 are 230.7, 229.4, 252.7, 278.8, 291.2, and 321.1 nm, respectively (Table1) and consistent with the trend observed from TEM (Figure 4). The particle size constantly increased during the sequential modification/conjugationmodification/conjugation steps steps fr fromom PMNP PMNP to to pPMNP-rtPA pPMNP-rtPA due due to to increasing increasing amount amount of of molecules molecules coatedcoated onon nanoparticlenanoparticle surface. surface. The The polydispersity polydispersity index index (PDI) (PDI) were allwere below all 0.22below (Table 0.221), (Table indicating 1), indicatingthat the particle that the size particle distribution size distribution is uniform is for uniform all samples, for all and samples, every sampleand every has sample good suspension has good suspensionstability [43 stability]. [43].

FigureFigure 5. 5. CharacterizationCharacterization of of Fe Fe3O3O4 4magneticmagnetic nanoparticles nanoparticles (MNP), (MNP), oleic oleic acid acid coated coated magnetic magnetic nanoparticlesnanoparticles (OMNP), (OMNP), PLGA PLGA magnetic magnetic nanoparticles nanoparticles (PMNP), (PMNP), avidin-conjugated avidin-conjugated PLGA PLGA magnetic magnetic nanoparticlesnanoparticles (PMNP-avidin), (PMNP-avidin), rtPA-conjugated rtPA-conjugated PLGA PLGA magnetic magnetic nanoparticles nanoparticles (PMNP-rtPA), (PMNP-rtPA), and and peptide/rtPA-conjugatedpeptide/rtPA-conjugated PLGA PLGA magnetic magnetic nanopartic nanoparticlesles (pPMNP-rtPA) (pPMNP-rtPA) by by dynamic dynamic light light scattering scattering (DLS)(DLS) ( (aa)) and and X-ray X-ray diffraction diffraction (XRD) (XRD) ( (bb).).

Table 1. The average particle size, polydispersity index (PDI), zeta potential, and crystal size of Table 1. The average particle size, polydispersity index (PDI), zeta potential, and crystal size of different nanoparticles. different nanoparticles. Average Diameter 2 Zeta Potential Crystal Size 3 Sample 1 PDI Zeta Potential Crystal Size 3 Sample 1 Average(nm) Diameter 2 (nm) PDI (mV) (nm) (mV) (nm) MNP 230.7 17.1 0.19 0.05 12.1 2.9 10.2 ± ± ± MNPOMNP 229.3230.718.5 ± 17.1 0.20 0.190.01 ± 0.05 17.6 12.1 ±0.9 2.9 10.2 10.5 ± ± − ± OMNPPMNP 252.7229.313.4 ± 18.5 0.19 0.200.04 ± 0.01 30.0−17.6 ±1.3 0.9 10.5 10.3 ± ± − ± PMNP-avidin 278.8 19.1 0.21 0.02 25.8 0.4 11.7 PMNP 252.7± ± 13.4 ± 0.19 ± 0.04 −−30.0± ± 1.3 10.3 PMNP-rtPA 291.2 27.3 0.21 0.04 24.4 1.1 10.7 ± ± − ± PMNP-avidinpPMNP-rtPA 321.1278.826.9 ± 19.1 0.22 0.210.05 ± 0.02 22.1−25.8 ±2.0 0.4 11.7 10.9 ± ± − ± 1 MNP:PMNP-rtPA iron oxide magnetic nanoparticles,291.2 ± 27.3 OMNP: oleic acid 0.21 coated ± 0.04 magnetic − nanoparticles,24.4 ± 1.1 PMNP: PLGA 10.7 magnetic nanoparticles, PMNP-avidin: avidin-conjugated PLGA magnetic nanoparticles, PMNP-rtPA: rtPA-conjugated PLGA magneticpPMNP-rtPA nanoparticles, pPMNP-rtPA:321.1 ± peptide 26.9 /rtPA-conjugated 0.22 ± PLGA 0.05 magnetic−22.1 nanoparticles. ± 2.0 2 Determined 10.9 from dynamic light scattering (DLS). 3 Determined from x-ray diffraction (XRD). 1 MNP: iron oxide magnetic nanoparticles, OMNP: oleic acid coated magnetic nanoparticles, PMNP: PLGA magnetic nanoparticles, PMNP-avidin: avidin-conjugated PLGA magnetic nanoparticles, From electrophoretic mobility measurements, the average zeta potential of MNP is positive at PMNP-rtPA: rtPA-conjugated PLGA magnetic nanoparticles, pPMNP-rtPA: peptide/rtPA- 12.1 mV, due to residual ammonium ions associated with the particles as ammonia was used for the conjugated PLGA magnetic nanoparticles. 2 Determined from dynamic light scattering (DLS). 3 co-precipitationDetermined from of ferric x-ray and diffraction ferrous (XRD). ions during the synthesis (Table1). This value changed to negative values for OMNP ( 17.6 mV) and PMNP ( 30 mV) due to the carboxylic group of oleic acid and − − PLGA. The zeta potential shifted to 25.8 mV after avidin modification as the isoelectric point (pI) From electrophoretic mobility measurements,− the average zeta potential of MNP is positive at 12.1of this mV, protein due to is residual approximately ammonium 10 [44 ions]. A associated similar eff withect due the toparticles the positive as ammonia change was of rtPA used (pI for= the7.6) and peptide (pI =10) slightly changed the zeta potential to 24.4 mV (PMNP-rtPA) and 22.1 mV co-precipitation of ferric and ferrous ions during the synthesis− (Table 1). This value changed− to negative(pPMNP-rtPA). values for OMNP (−17.6 mV) and PMNP (−30 mV) due to the carboxylic group of oleic acid and PLGA.The x-ray The di zetaffraction potential (XRD) shifted patterns to − that25.8 are mV shown after inavidin Figure modification5b reveal six as characteristic the isoelectric peaks point at (pI)2θ = of30.3 this◦ ,protein 35.7◦, 43.4 is approximately◦, 53.87◦, 57.4◦ ,10 and [44]. 63.1 A◦ similarfor all samples,effect due which to the represent positive thechange (220), of (311),rtPA (pI (400), = 7.6)(422), and (511), peptide and (440)(pI =10) planes slightly of a changed cubic cells the and zeta corresponds potential to to − the24.4 magnetite mV (PMNP-rtPA) structure and (JCPDS −22.1 file mV no. (pPMNP-rtPA).82-1533). This confirms that all of the nanoparticles were pure Fe3O4 with a spinel structure and PLGA encapsulation or surface conjugation steps did not lead to phase change of Fe3O4. The average crystal size calculated from the Scherrer equation, using the highest diffraction peak at 2θ = 35.7◦, with the Int. J. Mol. Sci. 2020, 21, 2690 9 of 24

The x-ray diffraction (XRD) patterns that are shown in Figure 5b reveal six characteristic peaks at 2θ = 30.3°, 35.7°, 43.4°, 53.87°, 57.4°, and 63.1° for all samples, which represent the (220), (311), (400), (422), (511), and (440) planes of a cubic cells and corresponds to the magnetite structure (JCPDS file no. 82-1533). This confirms that all of the nanoparticles were pure Fe3O4 with a spinel structure and PLGA encapsulation or surface conjugation steps did not lead to phase change of Fe3O4. The average Int. J. Mol. Sci. 2020, 21, 2690 9 of 23 crystal size calculated from the Scherrer equation, using the highest diffraction peak at 2θ = 35.7°, with the XRD line broadening and assuming spherical crystals, ranged from 10 to 12 nm (Table 1), supportingXRD line broadeningthe results from and assuming TEM. spherical crystals, ranged from 10 to 12 nm (Table1), supporting theThe results Fourier from TEM.transform infrared (FTIR) spectra are shown in Figure 6a for all samples. The The Fourier transform infrared (FTIR) spectra are shown in Figure6a for all samples. characteristic peak of Fe3O4 MNP at 640 and 570 cm−1 represented the Fe–O bond while the peak in 1 theThe 3200 characteristic to 3600 cm− peak1 range of Feand3O at4 MNP 1650 atcm 640−1 might and 570 represent cm− represented the N–H bond the Fe–O contributed bond while from the residual peak 1 1 ammonia,in the 3200 as observed to 3600 cm from− range its positive and at 1650zeta potential cm− might (Table represent 1). For the OMNP, N–H bondthe characteristic contributed peak from at residual ammonia, as observed from its positive zeta potential (Table1). For OMNP, the characteristic 2950, 2900, and 2850 cm−1 represented the C–H bond of –CH, –CH2, and –CH3, respectively. The peak at 2950, 2900, and 2850 cm 1 represented the C–H bond of –CH, –CH , and –CH , respectively. characteristic peak that ranged 1750–1680− cm−1 represented the C=O (carboxylic2 acid) 3bond, while the The characteristic peak that ranged 1750–1680 cm 1 represented the C=O (carboxylic acid) bond, characteristic peak ranged 1450–1400 cm−1 represented− the O–H (carboxylic acid) bond and the C–H while the characteristic peak ranged 1450–1400 cm 1 represented the O–H (carboxylic acid) bond and − −1 −1 bond, supporting oleic acid coating. For PMNP, the characteristic peak at 1180 cm and 10801 cm the C–H bond, supporting oleic acid coating. For PMNP, the characteristic peak at 1180 cm− and represented1 the C–O bond of ester and alcohol, indicating the successful entrapment of OMNP in 1080 cm− represented the C–O bond of ester and alcohol, indicating the successful entrapment of PLGA. A new characteristic peak at 1350 cm−1 that was assigned1 to the C–N (acryl) bond was observed OMNP in PLGA. A new characteristic peak at 1350 cm− that was assigned to the C–N (acryl) bond forwas PMNP-avidin, observed for PMNP-avidin,supporting surface supporting modification surface modification of PMNP with of PMNP avidin. with No avidin. new Nopeaks new could peaks be identifiedcould be after identified peptide after and peptide rtPA andconjugation, rtPA conjugation, as expected. as expected.

FigureFigure 6. 6.CharacterizationCharacterization of of Fe Fe3O3O4 4magneticmagnetic nanoparticles nanoparticles (MNP),(MNP), oleicoleic acidacid coated coated magnetic magnetic nanoparticlesnanoparticles (OMNP), (OMNP), PLGA PLGA coated coated magnetic nanoparticlesnanoparticles (PMNP),(PMNP), avidin-conjugated avidin-conjugated PLGA PLGA coatedcoated magnetic nanoparticlesnanoparticles (PMNP-Avidin), (PMNP-Avidin), rtPA-conjugated rtPA-conjugated PLGA coated PLGA magnetic coated nanoparticles magnetic nanoparticles(PMNP-rtPA) (PMNP-rtPA) and peptide/rtPA-conjugated and peptide/rtPA-conjugated PLGA magnetic PLGA nanoparticles magnetic (pPMNP-rtPA) nanoparticles by (pPMNP- Fourier transform infrared (FTIR) spectroscopy (a) and thermogravimetric analysis (TGA) (b). rtPA) by Fourier transform infrared (FTIR) spectroscopy (a) and thermogravimetric analysis (TGA) (b).The thermogravimetric analysis (TGA) results that are shown in Figure6b reveal weight loss starting from 270 ◦C for OMNP compared with MNP due to the decomposition of the organic oleic acid The thermogravimetric analysis (TGA) results that are shown in Figure 6b reveal weight loss molecules on Fe3O4 MNP and the final residual weight at 700 ◦C indicates 10% (w/w) oleic acid in OMNP starting from 270 °C for OMNP compared with MNP due to the decomposition of the organic oleic (Table2). For PLGA-encapsulated nanoparticles, substantial weight loss near 275 ◦C was evident due to acidPLGA molecules decomposition on Fe3O and4 MNP the residualand the weights final residual of PMNP, weight PMNP-avidin, at 700 °C PMNP-rtPA, indicates 10% and ( pPMNP-rtPAw/w) oleic acid in OMNP (Table 2). For PLGA-encapsulated nanoparticles, substantial weight loss near 275 °C was at 700 ◦C were approximately 14.0%, 13.4%, 17.1%, and 16.9%, respectively (Table2). For comparison, evidentthe Fe 3dueO4 content to PLGA that decomposition was determined and from the inductively residual weights coupled of plasma-optical PMNP, PMNP-avidin, emission spectrometry PMNP-rtPA, and(ICP-OES) pPMNP-rtPA for MNP, at 700 OMNP, °C were PMNP, approximately PMNP-avidin, 14.0%, PMNP-rtPA, 13.4%, 17.1%, and pPMNP-rtPA and 16.9%, respectively are 97.8%, 87.1%, (Table 2).13.5%, For comparison, 12.3%, 10.5%, the and Fe3O 10.1%,4 content respectively that was (Tabledetermined2). Separate from experimentinductively indicatescoupled plasma-optical the residual emissionweight ofspectrometry PLGA, avidin, (ICP-OES) and rtPA afterfor MNP, burning OMNP, to 700 ◦CPMNP, in nitrogen PMNP-avidin, were 0%, 20.1%, PMNP-rtPA, and 33.7%, and pPMNP-rtPArespectively, are due 97.8%, to their 87.1%, respective 13.5%, synthetic 12.3%, and 10.5%, natural and material 10.1%, characteristicsrespectively (Table (data not 2). shown). Separate experimentWhen considering indicates the the residual residual weight weight of avidin of PLGA, and rtPA, avidin, the Fe and3O4 rtPAcontent after from burning TGA and to ICP-OES 700 °C in nitrogenexperiments were could0%, 20.1%, be deemed and to33.7%, be consistent. respectively, due to their respective synthetic and natural material characteristics (data not shown). When considering the residual weight of avidin and rtPA, the Fe3O4 content from TGA and ICP-OES experiments could be deemed to be consistent.

Int. J. Mol. Sci. 2020, 21, 2690 10 of 23

Table 2. The residual weight at 700 ◦C from thermogravimetric analysis (TGA) and the Fe3O4 content determined from inductively coupled plasma-optical emission spectrometry (ICP-OES) and superconducting quantum interference device (SQUID).

Residual Weight from TGA Fe O from ICP-OES Fe O from SQUID Sample 1 3 4 3 4 (%) (%) (%) MNP 97.9 97.8 1.6 100.0 0.05 ± ± OMNP 87.7 87.1 2.1 90.9 0.11 ± ± PMNP 14.0 13.5 0.3 15.5 0.06 ± ± PMNP-avidin 13.4 12.3 0.8 14.4 0.11 ± ± PMNP-rtPA 17.1 10.5 0.3 13.3 0.07 ± ± pPMNP-rtPA 16.9 10.1 0.4 13.5 0.07 ± ± 1 MNP: iron oxide magnetic nanoparticles, OMNP: oleic acid coated magnetic nanoparticles, PMNP: PLGA magnetic nanoparticles, PMNP-avidin: avidin-conjugated PLGA magnetic nanoparticles, PMNP-rtPA: rtPA-conjugated PLGA magnetic nanoparticles, pPMNP-rtPA: peptide/rtPA-conjugated PLGA magnetic nanoparticles.

Figure7a shows the magnetization curves of all samples from superconducting quantum interference device (SQUID) analysis at room temperature. The remnant (residue) magnetization (without magnetic field) was determined from Figure7b to determine the superparamagnetic property. The remnant magnetization of MNP is close to zero (0.21 emu/g), endorsing its superparamagnetic property and consistent with its ~10 nm size from TEM and XRD experiments and in accordance with reported extremely low or non-existent remanence of Fe3O4 NMP [45]. The remnant magnetization values of OMNP, PMNP, PMNP-avidin, PMNP-rtPA, and pPMNP-rtPA are also close to zero (< 0.3 emu/g) (Table2), indicating MNP retained the characteristic superparamagnetism behavior during the synthesis of PLGA-based nanoparticles. The superparamagnetism behavior refers that all magnetic nano-carriers can be magnetized when an external magnetic field is applied and there is no residual magnetic interaction after removing the magnetic field. This behavior depends on the size of iron oxide MNP and it generally occurs below ~20 nm particle size, which supports the characterizing results from TEM and XRD. Undoubtedly, the use of superparamagnetic pPMNP-rtPA confers an important property for in vivo application as the drug-conjugated magnetic nano-carrier could be easily dispersed after guidance to the target site by removing the applied magnetic field, which can prevent undesirable particle agglomeration and possible vessel blockage [46]. The saturation magnetization value of OMNP reduced to 90% of that of MNP, consistent with the trend of MNP content from TGA or ICP-OES. This value reduced drastically to ~15% of that of MNP for all PLGA-encapsulated nanoparticles (i.e., PMNP, PMNP-avidin, PMNP-rtPA, and pPMNP-rtPA), which is due to the reduced MNP weight in the nanoparticles. Although the magnetization value is deemed to be sufficient for magnetic guidance in magnetically targeted thrombolytic drug delivery [47], dual targeting through synergistic ligand targeting will undoubtedly improve the drug targeting and clot lysis efficacy.

2.3. In Vitro Biocompatibility The biocompatibility was first determined from the cytocompatibility of PMNP at different concentrations after being incubated with NIH 3T3 fibroblasts for 24 and 48 h. The relative cell viability after contacting with PMNP shows no significance difference from that of control (without PMNP) up to 0.2 mg/mL, indicating PMNP is not toxic, even at concentrations suitable for in vivo injection, as shown from Figure8a. Using this concentration, we further tested the cytotoxicity of all nanoparticles during pPMNP-rtPA synthesis. The relative cell viability of all samples not only remained above 70% at both time points, to conform with the cytocompatibility standards of ISO 10993-5, no significance difference was found for all groups when compared to the control (PBS), as shown in Figure8b. Int. J. Mol. Sci. 2020, 21, 2690 11 of 23 Int. J. Mol. Sci. 2020, 21, 2690 11 of 24

Figure 7.7. CharacterizationCharacterization of of Fe 33OO44 magneticmagnetic nanoparticles nanoparticles (MNP), (MNP), oleic oleic acid coated magnetic nanoparticles (OMNP), PLGA coated magnetic nanoparticles (PMNP), avidin-conjugated PLGA coated magneticmagnetic nanoparticles nanoparticles (PMNP-avidin), (PMNP-avidin), rtPA-conjugated rtPA-conjugated PLGA coated PLGA magnetic coated nanoparticles magnetic (PMNP-rtPA),nanoparticles (PMNP-rtPA), and peptide /andrtPA-conjugated peptide/rtPA-con PLGAjugated magnetic PLGA nanoparticlesmagnetic nanoparticles (pPMNP-rtPA) (pPMNP- by superconductingrtPA) by superconducting quantum interferencequantum interference device (SQUID) device hysteretic (SQUID) magnetization hysteretic magnetization curves (a). The curves remnant (a). Int. J. Mol. Sci. 2020, 21, 2690 12 of 24 (residue)The remnant magnetization (residue) magnetization of all samples of is all close samples to zero, is asclose shown to zero, in (b as). shown in (b).

2.3. In Vitro Biocompatibility The biocompatibility was first determined from the cytocompatibility of PMNP at different concentrations after being incubated with NIH 3T3 fibroblasts for 24 and 48 h. The relative cell viability after contacting with PMNP shows no significance difference from that of control (without PMNP) up to 0.2 mg/mL, indicating PMNP is not toxic, even at concentrations suitable for in vivo injection, as shown from Figure 8a. Using this concentration, we further tested the cytotoxicity of all nanoparticles during pPMNP-rtPA synthesis. The relative cell viability of all samples not only remained above 70% at both time points, to conform with the cytocompatibility standards of ISO 10993-5, no significance difference was found for all groups when compared to the control (PBS), as shown in Figure 8b. Other than cytocompatibility, further biocompatibility consideration is through hemocompatibility to determine whether nanoparticles may induce hemolysis when administrated through intravenous injection. After incubation with purified rat red blood cells diluted in PBS at 37 °C for 2 h, the absorption spectra of the supernatant of MNP, pPMNP, PMNP-rtPA, and pPMNP- rtPA was the same as PBS (negative control), but not water (positive control), due to busting of red blood cells and release of oxyhemoglobin to show absorbance maxima at 540 and 577 nm (Figure 8c). The hemolysis effect was further examined through a quantitative comparison of the hemolysis ratio by using the difference in OD540 between the sample and the negative control (PBS) divided by the difference between the positive control (water) and the negative control. No visible hemolysis was observed between all samples and PBS, but not water, and the hemolysis ratio was lower than 2% for every nanoparticle (Figure 8d), indicating the good hemocompatibility of the nanoparticles synthesized according to the hemocompatibility standards of ASTM F756–08. Taken together, the good cytocompatibility and negligible hemolytic activity of pPMNP-rtPA underlines its excellent biocompatibilityFigureFigure 8. 8.Cytocompatibility Cytocompatibility for in vivo application of of PMNP PMNP at [48].at didifferentff erent concentrationsconcentrations (n.s.: (n.s.: not not significant significant compared compared to to0 0 µgμ/mLg/mL at at 24 24 or or 48 48 h, h, all all samples) samples) ( a(a)) and and of of didifferentfferent nanoparticlesnanoparticles at at 200 200 μµg/mLg/mL (n.s.: (n.s.: not not significant significant comparedcompared to to PBS PBS at at 24 24 or or 48 48 h, h, all all samples)samples) ((b)) byby MTT assay assay after after contacting contacting with with NIH NIH 3T3 3T3 cells. cells. HemocompatibilityHemocompatibility of diofff erentdifferent nanoparticles nanoparticles was determinedwas determined from thefrom hemolysis the hemolysis assay by assay incubation by

withincubation diluted redwith blood diluted cells red in blood PBS atcells 37 in◦C PBS for at 2 37 hto °C obtain for 2 h theto obtain full-wavelength the full-wavelength absorption absorption spectra of thespectra supernatant of the supernatant (c) and the hemolysis(c) and the hemolysis ratio from ratio OD540 from(d ).OD Water540 (d). and Water PBS and were PBS used were as used the positiveas the andpositive the negative and the controls, negative respectively. controls, respectively.

2.4. Targeting Effects The magnetic targeting effect was examined by magnetic guiding experiments by visually observing the solution turbidity after placing a magnet at the side of a tube containing 1 mg/mL nanoparticle for 5 min. From Figure 9a, the time-lapsed change of supernatant turbidity indicated a

magnetic field that was created by using a 1700 gauss magnet was effective to guide magnetic nanoparticles synthesized in this study to tube side against the gravity force, which reveals the possibility of using pPMNP-rtPA for magnetically guided thrombolysis. We used pre-synthesized Cy 5.5-encapsulated magnetic nanoparticles to demonstrate the fibrin binding ability of pPMNP and the ligand targeting effect. The fluorescently labelled nanoparticles were placed in a fibrin plate, shaken for 10 min., washed, and examined under an inverted fluorescence microscope to reveal the difference in binding ability between PMNP and pPMNP toward fibrin. In contrast to control with PBS showing no background fluorescence, Cy 5.5-labelled PMNP showed minimum fluorescence signal due to non-specific binding, in contrast to abundant scattered fluorescence associated with pPMNP due to specific peptide-fibrin biding, as shown in Figure 9b. The peptide GPRPPGGSKGC used here has been conjugated to fluorescently labeled MNP and showed high affinities for thrombi for multimodal fluorescence and magnetic resonance imaging of intravascular thrombus [49]. Therefore, the peptide-conjugated magnetic nanoparticles (pPMNP) offer the possibility for ligand- mediated fibrin-targeted thrombolytic drug delivery to induce thrombolysis. When compared with using monoclonal antibody against fibrin as a targeting ligand, the peptides are smaller in size, easier

Int. J. Mol. Sci. 2020, 21, 2690 12 of 23

Other than cytocompatibility, further biocompatibility consideration is through hemocompatibility to determine whether nanoparticles may induce hemolysis when administrated through intravenous injection. After incubation with purified rat red blood cells diluted in PBS at 37 ◦C for 2 h, the absorption spectra of the supernatant of MNP, pPMNP, PMNP-rtPA, and pPMNP-rtPA was the same as PBS (negative control), but not water (positive control), due to busting of red blood cells and release of oxyhemoglobin to show absorbance maxima at 540 and 577 nm (Figure8c). The hemolysis e ffect was further examined through a quantitative comparison of the hemolysis ratio by using the difference in OD540 between the sample and the negative control (PBS) divided by the difference between the positive control (water) and the negative control. No visible hemolysis was observed between all samples and PBS, but not water, and the hemolysis ratio was lower than 2% for every nanoparticle (Figure8d), indicating the good hemocompatibility of the nanoparticles synthesized according to the hemocompatibility standards of ASTM F756–08. Taken together, the good cytocompatibility and negligible hemolytic activity of pPMNP-rtPA underlines its excellent biocompatibility for in vivo application [48].

2.4. Targeting Effects The magnetic targeting effect was examined by magnetic guiding experiments by visually observing the solution turbidity after placing a magnet at the side of a tube containing 1 mg/mL nanoparticle for 5 min From Figure9a, the time-lapsed change of supernatant turbidity indicated a magnetic field that was created by using a 1700 gauss magnet was effective to guide magnetic nanoparticles synthesized in this study to tube side against the gravity force, which reveals the possibility of using pPMNP-rtPA for magnetically guided thrombolysis. We used pre-synthesized Cy 5.5-encapsulated magnetic nanoparticles to demonstrate the fibrin binding ability of pPMNP and the ligand targeting effect. The fluorescently labelled nanoparticles were placed in a fibrin plate, shaken for 10 min, washed, and examined under an inverted fluorescence microscope to reveal the difference in binding ability between PMNP and pPMNP toward fibrin. In contrast to control with PBS showing no background fluorescence, Cy 5.5-labelled PMNP showed minimum fluorescence signal due to non-specific binding, in contrast to abundant scattered fluorescence associated with pPMNP due to specific peptide-fibrin biding, as shown in Figure9b. The peptide GPRPPGGSKGC used here has been conjugated to fluorescently labeled MNP and showed high affinities for thrombi for multimodal fluorescence and magnetic resonance imaging of intravascular thrombus [49]. Therefore, the peptide-conjugated magnetic nanoparticles (pPMNP) offer the possibility for ligand-mediated fibrin-targeted thrombolytic drug delivery to induce thrombolysis. When compared with using Int. J. Mol. Sci. 2020, 21, 2690 13 of 24 monoclonal antibody against fibrin as a targeting ligand, the peptides are smaller in size, easier to produceto produce than monoclonal than monoclonal antibodies, antibodies, less less likely likely to induce to induce any any immunologic immunologic reaction, reaction, butbut stillstill in most cases,most show cases, high show specificity high specif andicity binding and binding constants constants as monoclonal as monoclonal antibodies antibodies [50 [50].].

FigureFigure 9. (9.a )(a Magnetic) Magnetic fieldfield guidingguiding efficiency efficiency of ofdifferent different magnetic magnetic nanoparticles nanoparticles by attaching by attaching a a magnetmagnet (1700 (1700 gauss) to to the the side side of a tube of a contai tubening containing 1 mg/mL 1nanoparticle mg/mL nanoparticle in PBS for 5 min. in PBS(b) Fibrin- for 5 min. (b) Fibrin-bindingbinding efficiency effi ofciency PLGA of magnetic PLGA magneticnanoparticles nanoparticles (PMNP) and (PMNP) peptide-conjugated and peptide-conjugated PLGA magnetic PLGA magneticnanoparticles nanoparticles (pPMNP) (pPMNP) was determined was determined by contacti by contactingng Cy 5.5-labelled Cy 5.5-labelled magnetic magneticnanoparticles nanoparticles (0.1 (0.1 mgmg/mL)/mL) with with a afibrin fibrin plate plate for 10 for min. 10 min.The fluore Thescence fluorescence signal from signal the fromremnant the nanoparticles remnant nanoparticles bound to fibrin was observed under an inverted fluorescence microscope after washing (bar = 50 um). bound to fibrin was observed under an inverted fluorescence microscope after washing (bar = 50 um). 2.5. In Vitro Thrombolysis Static in vitro thrombolysis in vertical position, using a blood clot made from rat whole blood, was performed with or without a magnet, to determine the effectiveness of magnetic targeting in blood clot lysis (Figure 10a) [51]. For this purpose, the extent of clot lysis when pre-formed blood clot was treated by rtPA and PMNP-rtPA in a sample tube was compared under the influence of downward magnetic pulling force generated from a magnet placed at the bottom of the tube. The clots treated in PBS and PMNP were taken as controls. Blood clot lysis was monitored spectrophotometrically by measuring the absorbance of the supernatant at 415 nm (OD415) after 10 min. In PBS or PMNP without rtPA, the blood clot remains intact with the solution transparency similar to 0 min., as observed from Figure 10b. Nonetheless, the contact with rtPA or PMNP-rtPA resulted in red-colored supernatant solution due to release of red blood cell residues from the tight fibrin network in the blood clot. By comparing the OD415 values presented in Figure 10c, there is no significant difference in OD415 for rtPA and PMNP-rtPA without magnet, as observed before from the agar plate clot lysis assays (Figure 3c). Most importantly, only PMNP-rtPA, but not rtPA, shows magnet-sensitive lysis activity. The OD415 increased significantly (1.83 folds) due to a downward magnetic force being generated by the magnet to attract PMNP-rtPA toward the blood clot for enhanced fibrinolysis.

Figure 10. The in vitro static clot lysis with or without magnetic guidance in vertical position was carried out as illustrated in (a). The solution appearance after blood clot lysis (b) and the solution absorbance measured at 415 nm (OD415) (c) after incubating the blood clot with PBS, PMNP, rtPA or PMNP-rtPA solution (25 μU rtPA). * p < 0.05 compared with pPMNP-rtPA without magnet.

Int. J. Mol. Sci. 2020, 21, 2690 13 of 24 to produce than monoclonal antibodies, less likely to induce any immunologic reaction, but still in most cases, show high specificity and binding constants as monoclonal antibodies [50].

Figure 9. (a) Magnetic field guiding efficiency of different magnetic nanoparticles by attaching a magnet (1700 gauss) to the side of a tube containing 1 mg/mL nanoparticle in PBS for 5 min. (b) Fibrin- binding efficiency of PLGA magnetic nanoparticles (PMNP) and peptide-conjugated PLGA magnetic nanoparticles (pPMNP) was determined by contacting Cy 5.5-labelled magnetic nanoparticles (0.1 mg/mL) with a fibrin plate for 10 min. The fluorescence signal from the remnant nanoparticles bound to fibrin was observed under an inverted fluorescence microscope after washing (bar = 50 um).

2.5. InInt. Vitro J. Mol. Thrombolysis Sci. 2020, 21, 2690 13 of 23 Static in vitro thrombolysis in vertical position, using a blood clot made from rat whole blood, was performed2.5. In Vitro Thrombolysiswith or without a magnet, to determine the effectiveness of magnetic targeting in blood clot lysis (Figure 10a) [51]. For this purpose, the extent of clot lysis when pre-formed blood clot Static in vitro thrombolysis in vertical position, using a blood clot made from rat whole blood, was treated by rtPA and PMNP-rtPA in a sample tube was compared under the influence of was performed with or without a magnet, to determine the effectiveness of magnetic targeting in blood downwardclot lysis magnetic (Figure 10 pullinga) [51]. Forforce this generated purpose, thefrom extent a magnet of clot lysisplaced when at pre-formedthe bottom blood of the clot tube. was The clots treatedtreated by in rtPA PBS and PMNP-rtPAand PMNP in were a sample taken tube wasas controls. compared underBlood theclot influence lysis ofwas downward monitored spectrophotometricallymagnetic pulling force by generatedmeasuring from the a absorbance magnet placed of atthe the supernatant bottom of the at tube.415 nm The (OD clots415 treated) after 10 min. inIn PBSPBS and or PMNPPMNP were without taken rtPA, as controls. the blood Blood clot clot remains lysis was monitoredintact with spectrophotometrically the solution transparency by similarmeasuring to 0 min., the as absorbance observed of from the supernatant Figure 10b. at 415No nmnetheless, (OD415) afterthe contact 10 min. Inwith PBS rtPA or PMNP or PMNP-rtPA without resultedrtPA, in the red-colored blood clot remainssupernatant intact solution with the solutiondue to release transparency of red similar blood to cell 0 min, residues as observed from the from tight fibrinFigure network 10b. in Nonetheless, the blood clot. the contactBy comparing with rtPA the or OD PMNP-rtPA415 values resulted presented in red-colored in Figure supernatant10c, there is no solution due to release of red blood cell residues from the tight fibrin network in the blood clot. By significant difference in OD415 for rtPA and PMNP-rtPA without magnet, as observed before from the comparing the OD values presented in Figure 10c, there is no significant difference in OD for rtPA agar plate clot lysis assays415 (Figure 3c). Most importantly, only PMNP-rtPA, but not415 rtPA, shows and PMNP-rtPA without magnet, as observed before from the agar plate clot lysis assays (Figure3c). magnet-sensitive lysis activity. The OD415 increased significantly (1.83 folds) due to a downward Most importantly, only PMNP-rtPA, but not rtPA, shows magnet-sensitive lysis activity. The OD415 magneticincreased force significantly being generated (1.83 folds) by duethe tomagnet a downward to attract magnetic PMNP-rtPA force being toward generated the by blood the magnet clot for enhancedto attract fibrinolysis. PMNP-rtPA toward the blood clot for enhanced fibrinolysis.

FigureFigure 10. The 10. Thein vitroin vitro staticstatic clot clot lysis lysis with with or or withou withoutt magnetic guidanceguidance in in vertical vertical position position was was carried out as illustrated in (a). The solution appearance after blood clot lysis (b) and the solution carried out as illustrated in (a). The solution appearance after blood clot lysis (b) and the solution absorbance measured at 415 nm (OD415)(c) after incubating the blood clot with PBS, PMNP, rtPA or absorbance measured at 415 nm (OD415) (c) after incubating the blood clot with PBS, PMNP, rtPA or PMNP-rtPA solution (25 µU rtPA). * p < 0.05 compared with pPMNP-rtPA without magnet. PMNP-rtPA solution (25 μU rtPA). * p < 0.05 compared with pPMNP-rtPA without magnet. Dynamic in vitro thrombolysis was employed to verify enhanced clot lysis due to the conjugated peptide ligand at 37 ◦C and 43 ◦C by slowly rotating the clot-containing vial (Figure 11a). The mixing of a blood clot with the nanocarrier was deemed to be necessary for showing the ligand targeting effect of pPMNP-rtPA originated from fibrin binding. As in static test, the color intensity of the red-colored supernatant solution increased only for rtPA, PMNP-rtPA, and pPMNP-rtPA with the thrombolytic action of rtPA releasing red blood cell residues from the clot (Figure 11b). Most importantly, a significant difference in OD415 was found between pPMNP-rtPA and pPMNP-rtPA, regardless of temperature (Figure 11c). This underscores the importance of using fibrin-avid peptide moiety to upregulate the clot lysis activity of immobilized rtPA, due to the preferential binding to fibrin (Figure9b). That temperature-dependent lysis was only being shown in rtPA-containing samples further supports rtPA-induced clot lysis, with higher rtPA enzymatic activity at 43 ◦C. Int. J. Mol. Sci. 2020, 21, 2690 14 of 24

Dynamic in vitro thrombolysis was employed to verify enhanced clot lysis due to the conjugated peptide ligand at 37 °C and 43 °C by slowly rotating the clot-containing vial (Figure 11a). The mixing of a blood clot with the nanocarrier was deemed to be necessary for showing the ligand targeting effect of pPMNP-rtPA originated from fibrin binding. As in static test, the color intensity of the red- colored supernatant solution increased only for rtPA, PMNP-rtPA, and pPMNP-rtPA with the thrombolytic action of rtPA releasing red blood cell residues from the clot (Figure 11b). Most importantly, a significant difference in OD415 was found between pPMNP-rtPA and pPMNP-rtPA, regardless of temperature (Figure 11c). This underscores the importance of using fibrin-avid peptide moiety to upregulate the clot lysis activity of immobilized rtPA, due to the preferential binding to Int. J.fibrin Mol. Sci.(Figure2020, 219b)., 2690 That temperature-dependent lysis was only being shown in rtPA-containing14 of 23 samples further supports rtPA-induced clot lysis, with higher rtPA enzymatic activity at 43 °C.

FigureFigure 11. 11.The Thein in vitro vitrodynamic dynamic clotclot lysis was was carried carried out out as asillustrated illustrated in (a in). The (a). solution The solution appearance appearance afterafter blood blood clot clot lysis lysis (b ()b and) and the the solution solution absorbance measured measured at at415 415 nm nm (OD (OD415) (415c) after)(c) incubating after incubating the bloodthe blood clot clot with with PBS, PBS, PMNP, PMNP, rtPA,rtPA, PMNP-rtPA, or or pPMNP-rtPA pPMNP-rtPA solution solution (50 μ (50U rtPAµU rtPAactivity). activity). * * # # p < 0.05p < 0.05 compared compared with with PMNP-rtPA PMNP-rtPA at 37 °C,◦C, p p< <0.050.05 compared compared with with PMNP-rtPA PMNP-rtPA at 43 °C. at43 ◦C.

AfterAfter thrombolysis thrombolysis testing testing in ain closed a closed system, system we, studiedwe studied simulated simulated vascular vascular embolization embolization induced by ainduced blood clotby a andblood pressure-driven clot and pressure-driven clot lysis clot in ly asis flow in a flow system system (Figure (Figure 12 12a).a). The The best pPMNP- pPMNP-rtPA rtPA sample was tested in the flow thrombolysis model at 37 °C while using a water jacketed glass sample was tested in the flow thrombolysis model at 37 ◦C while using a water jacketed glass tube. tube. A blood clot was placed at the bottom of a tube with reduced diameter and lodged tightly within A blood clot was placed at the bottom of a tube with reduced diameter and lodged tightly within the tube inner circumference, in order to restrict lysis from the clot surface, and the sample was theinjected tube inner from circumference, a side opening. A in magnet order towas restrict placed lysisbelow from the clot, the outside clot surface, of the tube and to the introduce sample was injectedmagnetic from guidance. a side opening. A fluid Apressure magnet gradient was placed was generated below the with clot, PBS outside flow from of the the tube top to at introduce 0.5 magneticmL/min. guidance. using a syringe A fluid pump. pressure At ti gradientme zero, a was different generated sample with was PBS introduced flow from into the the topflow at system 0.5 mL /min usingand a syringethe reperfusion pump. Attime time for zero,clot adissolution different sample(clot lysis was time) introduced was compared into the among flow systemdifferent and the reperfusiontreatment time groups for clotto determine dissolution the (clotlysis lysisefficiency. time) The was blood compared clot lysis among efficiency different of pPMNP-rtPA treatment groups to determinetreatment thewas lysis enhanced efficiency. with Thesignificantly blood clot reduced lysis eclotfficiency lysis time of pPMNP-rtPA (51 min.), when treatment reperfusion was and enhanced withdisengaging significantly of reducedclot from clotthe tube lysis occurs. time (51 This min), could when be compared reperfusion with andfree rtPA disengaging (86 min.) ofas clotshown from the in Figure 12b. The combination of magnetic and ligand targeting resulted in the shortest reperfusion tube occurs. This could be compared with free rtPA (86 min) as shown in Figure 12b. The combination time with 2.3-fold increase in lysis efficiency when compared to the pPMNP group. of magnetic and ligand targeting resulted in the shortest reperfusion time with 2.3-fold increase in lysisInt. effi J.ciency Mol. Sci. when2020, 21 compared, 2690 to the pPMNP group. 15 of 24

FigureFigure 12. 12.The The schematic schematic diagramdiagram of of a aflow flow model model to ev toaluate evaluate pressure-driven pressure-driven thrombolysis thrombolysis at 0.5 at 0.5 mLmL/min/min ((aa).). At time time zero, zero, different different sample sample (PBS, (PBS, pPMN pPMNP,P, rtPA, rtPA,or pPMNP-rtPA) or pPMNP-rtPA) was introduced was introduced into into thethe flowflow system system and and the the blood blood clot clot lysis lysis time time was wasrecorded recorded when whenreperfusion reperfusion occurs to occurs determine to determine the the lysislysis e efficiencyfficiency ( b).). * *pp << 0.050.05 compared compared with with rtPA. rtPA.

2.6. In Vivo Thrombolysis The clinical application potential of dual targeted delivery of rtPA was evaluated by determining the in vivo fibrinolytic efficacy of pPMNP-rtPA where induce targeted thrombolysis under magnetic guidance was assessed in a rat embolic model [48]. Five minutes after lodging the left iliac artery upstream of the pubic epigastric artery with a blood clot, pPMNP (vehicle), free rtPA, or pPMNP- rtPA was administered intra-arterially via the right iliac artery. Magnetic guidance with a 0.5 T magnet started immediately after drug administration. This was repeated in five successive cycles (5 min. each) from the branch of abdominal aorta and iliac artery to the branch of iliac artery and the femoral artery (1 min.) and to the lower extremities (4 min.) (Figure 13a). The blood flow pattern changes around the surrounding areas of the hind limb was acquired using a laser speckle contrast imager, while the rates of iliac blood flow (IBF) and aortic blood flow (ABF) were determined using ultrasonic flow probes. The hind limp perfusion reduced dramatically from the basal after introducing the clot, which created vascular thrombosis for evaluation of the fibrinolytic activity of rtPA in vivo, as shown in Figure 13b. It is evident that pPMNP-rtPA injection improved the blood perfusion in the hind limb 35 min. post-treatment due to restoring the blood flow downstream of the blood clot site, which was not achievable in both pPMNP and rtPA groups. The rtPA group showed increased blood perfusion only after 60 min. treatment to be similar to the pPMNP-rtPA, while the pPMNP group showed a similar thrombosis condition without any improvement of the perfusion rate (Figure 13b). The time- lapsed hemodynamics results, including ABF (Figure 13c) and IBF (Figure 13d), were shown after different treatments. Before clot lodging, the IBF values were 9.4 ± 1.0, 7.9 ± 0.9 and 10.1 ± 1.1 mL/min. for pPMNP, rtPA, and pPMNP-rtPA, respectively. Nonetheless, the IBF values reduced to nearly 0% for all groups after thromboembolism, due to the complete occlusion of the left iliac artery. Simultaneously, the ABF also reduced to 6 mL/min. in all groups. After the treatment, the IBF restored to 0.7 ± 0.3, 4.8 ± 0.8, and 4.9 ± 1.3 mL/min. for pPMNP, rtPA, and pPMNP-rtPA, which was 8.1 ± 3.5%, 60.7 ± 8.8%, and 46.9 ± 10.8% of the respective initial value. From 10 to 35 min., there was a significant difference in IBF between rtPA and pPMNP-rtPA, as revealed from the perfusion image (Figure 13b). However, there was no significant difference in IBF between rtPA and pPMNP-rtPA from 40 min. to 120 min., but both showed significant difference from pPMNP. In comparison, the ABF did not show dramatic recovery as IBF, although some recovery of ABF was also observed for the rtPA and pPMNP-rtPA groups after 100 min. due to downstream restoration of IBF. Most importantly, as the dosage of pPMNP-rtPA was 0.3 U/kg compared to 1.5 U/kg for rtPA, we

Int. J. Mol. Sci. 2020, 21, 2690 15 of 23

2.6. In Vivo Thrombolysis The clinical application potential of dual targeted delivery of rtPA was evaluated by determining theInt. J.in Mol. vivo Sci.fibrinolytic 2020, 21, 2690 e fficacy of pPMNP-rtPA where induce targeted thrombolysis under magnetic16 of 24 guidance was assessed in a rat embolic model [48]. Five minutes after lodging the left iliac artery upstreamdemonstrated of the comparable pubic epigastric efficacy artery in restoring with a blood blood clot, flow pPMNP and in (vehicle), vivo thrombolysis free rtPA, or using pPMNP-rtPA only 20% wasdosage administered of free rtPA. intra-arterially By allowing the via use the rightof a lowe iliacr artery.rtPA dosage Magnetic with guidance shortened with clot a lysis 0.5 T time, magnet this starteddual targeted immediately nanomedicine after drug approach administration. will be a This feasible was repeatedstrategy for in fiveimproving successive the cyclesefficacy (5 and min safety each) fromof rtPA the in branch reducing of abdominal bleeding aortarisk [52]. and iliacOverall, artery the to thein vivo branch study of iliac suggested artery andthe theclinical femoral feasibility artery (1using min) dual and targeted to the lower delivery extremities strategy (4 for min) rtPA (Figure and that 13a). dual The targeted blood flow pPMNP pattern is an changes efficient around functional the surroundingnano-carrier areasfor comparable of the hind in limb vivo was thrombolysis acquired using outcomes a laser speckleat reduced contrast rtPA imager, dosage, while which the might rates ofreduce iliac bloodthe hemorrhagic flow (IBF) and side aortic effects blood in clinical flow (ABF) thrombolytic were determined therapy. using ultrasonic flow probes.

FigureFigure 13.13. TheThe inin vivovivo thrombolyticthrombolytic eeffectsffects ofof pPMNP-rtPApPMNP-rtPA byby evaluatingevaluating targetedtargeted thrombolysisthrombolysis inin aa ratrat embolicembolic model. model. ( a(a)) A A blood blood clot clot was was lodged lodged into into the the left left iliac iliac artery artery and and after after 5 min 5 min. pPMNP pPMNP (n = (n5), = rtPA5), rtPA (1.5 (1.5 U/kg; U/kg;n = n5) = or5) pPMNP-rtPAor pPMNP-rtPA (0.3 (0.3 U/ U/kg;kg; n = n 5)= 5) was was delivered delivered from from the the right right iliac iliac artery arteryand and subjectsubject toto magnetic guidance guidance for for 25 25 min. min. (b ()b Representative) Representative images images of blood of blood perfusion perfusion rate ratein the in hind the hindlimb limbarea areaat different at different time timepoints points afterafter clot intr clotoduction introduction using using a laser a laser speckle speckle contrast contrast imager. imager. The Theabdominal abdominal aortic aortic blood blood flow flow (ABF) (ABF) (c) (cand) and iliac iliac blood blood flow flow (IBF) (IBF) ( (dd)) before before and afterafter bloodblood clotclot introductionintroduction (0(0 min)min.) was was determined determined from from ultrasonic ultrasonic flow flow probes. probes. **p p << 0.05 rtPA or pPMNP-rtPA vs.vs. PMNP,PMNP,# #p p< < 0.050.05 rtPArtPA vs.vs. pPMNP-rtPA.pPMNP-rtPA.

3. Materials and Methods

3.1. Materials 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), iron (II) chloride tetrahydrate, and iron(III) chloride hexahydrate were purchased from Organics, Thermo Fisher Scientific (Geel, Belgium). Acetone, dichloromethane (DCM), oleic acid,

Int. J. Mol. Sci. 2020, 21, 2690 16 of 23

The hind limp perfusion reduced dramatically from the basal after introducing the clot, which created vascular thrombosis for evaluation of the fibrinolytic activity of rtPA in vivo, as shown in Figure 13b. It is evident that pPMNP-rtPA injection improved the blood perfusion in the hind limb 35 min post-treatment due to restoring the blood flow downstream of the blood clot site, which was not achievable in both pPMNP and rtPA groups. The rtPA group showed increased blood perfusion only after 60 min treatment to be similar to the pPMNP-rtPA, while the pPMNP group showed a similar thrombosis condition without any improvement of the perfusion rate (Figure 13b). The time-lapsed hemodynamics results, including ABF (Figure 13c) and IBF (Figure 13d), were shown after different treatments. Before clot lodging, the IBF values were 9.4 1.0, 7.9 0.9 and 10.1 1.1 mL/min for ± ± ± pPMNP, rtPA, and pPMNP-rtPA, respectively. Nonetheless, the IBF values reduced to nearly 0% for all groups after thromboembolism, due to the complete occlusion of the left iliac artery. Simultaneously, the ABF also reduced to 6 mL/min in all groups. After the treatment, the IBF restored to 0.7 0.3, 4.8 ± ± 0.8, and 4.9 1.3 mL/min for pPMNP, rtPA, and pPMNP-rtPA, which was 8.1 3.5%, 60.7 8.8%, and ± ± ± 46.9 10.8% of the respective initial value. From 10 to 35 min, there was a significant difference in IBF ± between rtPA and pPMNP-rtPA, as revealed from the perfusion image (Figure 13b). However, there was no significant difference in IBF between rtPA and pPMNP-rtPA from 40 min to 120 min, but both showed significant difference from pPMNP. In comparison, the ABF did not show dramatic recovery as IBF, although some recovery of ABF was also observed for the rtPA and pPMNP-rtPA groups after 100 min due to downstream restoration of IBF. Most importantly, as the dosage of pPMNP-rtPA was 0.3 U/kg compared to 1.5 U/kg for rtPA, we demonstrated comparable efficacy in restoring blood flow and in vivo thrombolysis using only 20% dosage of free rtPA. By allowing the use of a lower rtPA dosage with shortened clot lysis time, this dual targeted nanomedicine approach will be a feasible strategy for improving the efficacy and safety of rtPA in reducing bleeding risk [52]. Overall, the in vivo study suggested the clinical feasibility using dual targeted delivery strategy for rtPA and that dual targeted pPMNP is an efficient functional nano-carrier for comparable in vivo thrombolysis outcomes at reduced rtPA dosage, which might reduce the hemorrhagic side effects in clinical thrombolytic therapy.

3. Materials and Methods

3.1. Materials 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), iron (II) chloride tetrahydrate, and iron(III) chloride hexahydrate were purchased from Organics, Thermo Fisher Scientific (Geel, Belgium). Acetone, dichloromethane (DCM), oleic acid, polyvinyl alcohol (PVA) (hydrolyzed, M.W. 30,000–70,000), and plasminogen activator inhibitor-1 (PAI-1) were purchased from Sigma-Aldrich (St Louis, MO, USA). Poly(lactic-co-glycolic acid) (PLGA) (lactide/glycolide = 50/50, molecular weight = 15,000~30,000, intrinsic viscosity = 0.35 0.05 dL/g) was ± purchased form Green Square Co. (Taoyuan, Taiwan). Recombinant tissue plasminogen activator (rtPA, Actilyse®) was provided form Boehringer Ingelheim (Ingelheim am Rhein, Germany). Avidin and biotin-PEG-maleimide (M.W. 3500) were purchased form Merck (Darmstadt, Germany) and JenKem Technology USA Inc. (Plano, TX, USA), respectively.

3.2. Preparation of Magnetic Nanoparticles The oleic acid coated iron oxide magnetic nanoparticles (OMNP) were prepared by the co-precipitation method. FeCl 6H O (2.15 g) and FeCl 4H O (0.79 g) were dissolved in 50 mL 3· 2 2· 2 ddH2O in a three necks glass reactor. Nitrogen was purged for 10 min to remove oxygen, followed by heating the solution to 60 ◦C. Five milliliter of NH4OH was added dropwise with a needle into the reactor and then reacted for 30 min under nitrogen by stirring at 500 rpm. The solution was cooled to room temperature while under continuous nitrogen purging. The magnetic nanoparticles was washed three times with distilled deionized water (ddH2O) after recovering with a magnet and the final solution pH was adjusted to 5 using 0.1 N HCl. The solution was sonicated for 10 min, Int. J. Mol. Sci. 2020, 21, 2690 17 of 23

followed by nitrogen purging for 10 min before heating to 60 ◦C. Ten milliliters of 2.77% (v/v) oleic acid solution prepared in acetone was then added dropwise with a needle into the reactor and reacted for 30 min under 500 rpm stirring. Nitrogen purge continued while cooling down the solution to room temperature. After removing acetone by evaporation, OMNPs was dispersed in chloroform to 1% (w/w) for storage. The OMNPs prepared above were entrapped in PLGA polymer matrix by the single emulsion-solvent evaporation method to prepare PLGA MNP (PMNP). Fifty milligram of PLGA was dissolved in 2.5 mL organic solvent mixture (2.25 mL acetone and 0.25 mL DCM), followed by sonicating for 10 min. Five hundred microliters of OMNPs prepared above (10 mg OMNPs in 1 mL chloroform) was added, vortexed for 1 min and sonicated for 1 min in a sonication bath. The mixture was quickly poured into 25 mL phosphate buffered saline (PBS), vortexed for 1 min and sonicated for 5 min (Q700, Qsonica, Newtown, CT, USA). The resulting solution containing PMNPs was added dropwise into 50 mL 0.3% (w/v) PVA solution prepared in PBS and stirred at 1000 rpm. After incubating at 30 ◦C for 12 h, the solution was placed in a rotary evaporator and then dried for 1 h at 30 ◦C for removing residual solvents and washed twice with ddH2O using magnetic decantation. To modify the surface of PMNPs with avidin, 1 mg/mL PMNPs in PBS (pH 7.4) was mixed with 0.5 mL EDC (0.5 mg/mL) and 0.5 mL NHS (0.5 mg/mL) prepared in 0.5 M pH 5 MES buffer and rotated at 6 rpm for 1 h at room temperature for the activation of the carboxyl groups in PLGA. After washing twice with PBS, activated PMNP was recovered by magnetic decantation and re-dispersed in 1 mL PBS, followed by adding 1 mL avidin (0.2 mg/mL) in PBS and mixed by rotation at 6 rpm for 3 h at room temperature. After washing twice with PBS, PMNP that was conjugated with avidin (PMNP-avidin) was recovered by magnetic decantation and re-dispersed in 1 mL PBS for immobilization of peptide/rtPA using biotin-PEG-peptide/biotin-PEG-rtPA through avidin-biotin interaction (Figure1). To prepare biotin-PEG-peptide, 0.65 µg biotin-PEG-maleimide was mixed with 0.21 µg fibrin-avid peptide (GPRPPGGSKGC) in 1 mL PBS and then mixed at 4 ◦C for 24 h. Similarly, 27 µg biotin-PEG-maleimide was mixed with 0.5 mg rtPA in 1 mL PBS at 4 ◦C for 24 h to obtain biotin-PEG-rtPA. The solutions prepared above were mixed successively with PMNP-avidin, first by adding biotin-PEG-peptide to PMNP-avidin and mixed for 24 h at 4 ◦C, followed by adding biotin-PEG-rtPA and mixed for another 24 h at 4 ◦C. Finally, the particles were washed twice with PBS to obtain peptide/rtPA conjugated PLGA magnetic nanoparticles (pPMNP-rtPA). As a control, the peptide conjugated PLGA magnetic nanoparticles (pPMNP) or rtPA conjugated PLGA magnetic nanoparticles (PMNP-rtPA) were also prepared without adding biotin-PEG-peptide or biotin-PEG-rtPA during synthesis (Figure1).

3.3. Determination of rtPA and Peptide Conjugation The amount of avidin loading on nanoparticles was measured by the Pierce™ BCA protein assay kit from Thermo Fisher Scientific (Geel, Belgium). All of the samples were collected after modification and dilute with PBS to appropriate concentration. Fifty microliter sample (1 mg/mL) and 1 mL BCA reagent were added into a 2 mL brown centrifuge tube and reacted for 30 min at 37 ◦C. After cooling to room temperature, the chromogenic product was measured by an UV-Vis spectrophotometer at 562 nm. The loading content of rtPA was determined as mg of protein per mg of nanoparticles. For rtPA loading, the activity of immobilized rtPA was spectrophotometrically determined from its amidolytic activity using S-2288 as a specific protease chromogenic substrate (Chromogenix, Mölndal, Sweden) following that manufacturer’s protocols. Two hundred microliter of samples (rtPA, PMNP-rtPA and pPMNP-rtPA) was diluted with 0.2 mL 0.1 M pH 7.4 Tris buffer and incubated at 37 ◦C for 5 min. Two hundred microliter of S-2288 substrate (1 mM) was added and vortexed immediately before incubating at 37 ◦C. To stop the enzymatic reaction, 0.1 mL acetic acid (20% (v/v)) was added after 30 s and solution absorbance was determined for the chromogenic product using an UV-Vis spectrophotometer at 405 nm (OD ) after magnetic separation. The amidolytic activity (U) was calculated as OD 405 405 × 0.313 0.2/0.5 using respective nanoparticles as the blank. × Int. J. Mol. Sci. 2020, 21, 2690 18 of 23

3.4. Charactrization of Physico-Chemical Properties The particle morphology of the prepared sample was analyzed using a transmission electron microscope (TEM, JEOL JEM2000 EXII, Tokyo, Japan) at an accelerating voltage of 100 kV. For OMNP, a diluted sample that was prepared in ethanol was used while other samples were prepared in ddH2O. After dropping a sample solution on the surface of cooper grids, it was dried in a 37 ◦C oven for 24 h and blotted with a filter paper. The sample was stained with 1% (w/v) phosphotungstic acid aqueous solution for 1 min before analysis. The particles size distribution was analyzed using a Zetasizer (Nano ZS 90, Malvern Instruments, Malvern UK) at 25 ◦C. The measurement was carried out by setting equilibration time at 30 s, measurement duration at automatic, number of measurement at 3, and delay between measurements at 5 s. The OMNP sample was prepared by dispersing in ethanol at 0.1 mg/mL, while other samples were prepared by dispersing in ddH2O. The thermal property of the prepared sample was analyzed using a Q50 thermogravimetric analyzer (TGA) from TA Instruments (New Castle, DE, USA) while using a platinum pan and then purged with nitrogen at 60 mL/min. The temperature was raised from room temperature to 700 ◦C at a heating rate of 10 ◦C/min. All of the samples were dried in a 37 ◦C oven for 24 h before analysis. The element content of the prepared sample was analyzed using an inductively coupled plasma optical emission spectrometer (ICP-OES, 710-ES, Varian Inc., Palo Alto, CA, USA). All of the samples were digested in 37% HCl for 2 h at 60 ◦C and diluted with ddH2O into 15 mL before filtration through a 0.22 µm PVDF filter. The chemical structure of prepared sample was analyzed using X-ray diffraction (XRD, D2 PHASER, Bruker, MA, USA) using Cu-Kα radiation (λ = 1.54060 Å). The 2θ range was from 10 to 70◦ with 0.02◦ increment. All of the samples were dried in a 37 ◦C oven for 24 h before measurements. The crystalline size was calculated by the Debye–Scherrer equation.

k λ d = × (1) β cosθ × In Equation (1), d is the mean size of the crystalline domains, k is a dimensionless shape factor (0.9), λ is the X-ray wavelength, β is the line broadening at half the maximum intensity, and θ is the Bragg angle. The magnetic moment of prepared sample was analyzed using a superconducting quantum interference device (SQUID) magnetometer (MPMS XL-7, Quantum Design, San Diego, CA, USA). The magnetic field-dependent magnetization was measured from 10,000 to 10,000 Oe at 298 K. All of − ◦ the samples were dried in a 37 ◦C oven for 24 h and 40–50 mg of samples was analyzed. The magnetic field guidance of prepared samples was tested in PBS by attaching a magnet (1700 gauss) at side of a tube containing 1 mg/mL nanoparticle for 5 min.

3.5. Fibrin Binding Efficiency and Fibrinolysis Asaasy An agar solution was prepared by dissolving 0.2 g low-melting agar in 10 mL pH 7.4 Tris buffer and heated in a microwave. To induce clotting, 5 mg fibrinogen in 1 mL pH 7.4 Tris buffer, 10 mL clotting factor solution (1.8% NaCl and 0.3% CaCl2), and 5 mL thrombin solution (25 U) were added to the agar solution under stirring. For fibrin binding efficiency, 100 µL solution prepared above was added into a 96-well cell culture plate and incubated at 4 ◦C for 30 min to form solidified fibrin gel. To prepare fluorescently labelled nanoparticles, 5 µl Cy 5.5 (1 mg/mL) in dimethyl sulfoxide was added to the organic solvent mixture during PMNP and pPMNP preparation. For fibrin binding ability, 0.2 mL of test sample (PBS, 0.1 mg/mL PMNP or 0.1 mg/mL pPMNP) was added to each well and shaken for 10 min. After washing twice with PBS, microscopic images were taken using an inverted fluorescence microscope (Olympus IX71, Tokyo, Japan). For the agar plate fibrinolysis assay, the solution prepared above was poured into a 15 cm cell culture dish and spread evenly to obtain a homogeneous gel, which was then placed in a 4 ◦C refrigerator and incubated for 30 min to form solidified fibrin gel. After making nine 4-mm diameter holes on the agar plate as sample reservoirs, 10 µL of sample (20 µU rtPA) and 15 µL plasminogen Int. J. Mol. Sci. 2020, 21, 2690 19 of 23

prepared in PBS (1 mg/mL) were added into each hole and then incubated for 24 h at 37 ◦C to induce fibrinolysis. The fibrinolytic activity was determined by measuring the thickness of the clear zone around the sample hole well induced by the fibrinolytic activity of rtPA. The dose-dependent sensitivity of the test was confirmed from the same agar plate assay with 10 to 30 µU free rtPA following the same procedure. To study the inhibition of rtPA by PAI-1, 50 µM rtPA (free rtPA, PMNP-rtPA, or pPMNP-rtPA) and 50 µM PAI-1 were mixed in 40 µL 0.1 M Tris buffer (pH 7.4). After incubation at 37 ◦C for different times, 20 µL Tris buffer and 20 µL S-2288 substrate (1 mM) were added and incubated for another 30 min at 37 ◦C, followed by adding 20 µL acetic acid (20% (v/v)) to stop the reaction. The amidolytic activity was calculated from the solution absorbance at 405 nm after magnetic separation using an ELISA reader, as described above.

3.6. Determination of Thrombolysis Using Rotational Thromboelastometry Thromboelastography is a method to test blood coagulation efficiency by measuring the viscoelastic properties of developing whole blood clot and used in this study for in vitro thrombolysis assays from the percentage of clot that has lysed at different time points after rtPA treatment [53]. The whole blood that was obtained from Sprague-Dawley (SD) rats using cardiac puncture was citrated and incubated at 37 ◦C for 30 min before use. Thrombus formation was initiated by adding 300 µL whole blood, 20 µL PBS or rtPA sample (28.9 µU/mL), and 20 µL 12 mM CaCl2 in the sample cup, after which a thromboelastogram could be recorded with time using rotational thromboelastometry (ROTEM, Delta2000, Tem Innovations GmbH, Munich, Germany). The clot lysis index at different time points was calculated by dividing the amplitude at a given time with the maximum amplitude corresponding to the maximum clot firmness from the thromboelastogram to evaluate the thrombolysis effect that is induced by rtPA in vitro.

3.7. In Vitro Biocompatibility and Hemocompatibility To test biocompatibility, the in vitro cytotoxicity of nanoparticles was determined using NIH 3T3 cells and MTT assays. Briefly, 1 104 NIH 3T3 cells were seeded in a 96-well cell culture plate and then × cultured in DMEM medium supplemented with 10% fetal bovine serum at 37 ◦C in 5% CO2 atmosphere. After 24 h, the medium was removed and replenished with samples solution containing nanoparticles prepared in fresh medium and cultured for 24 and 48 h before measuring the mitochondria activity by the MTT assay. For MTT assay, the medium was removed and 100 µL MTT solution (10 µL 5mg/mL MTT reagent diluted with 90 µL medium) was added to each well and then incubated at 37 ◦C in 5% CO2 for 2 h. After dissolving the formazan crystals with dimethyl sulfoxide, the solution absorbance in each well was determined while using a microplate reader (Synergy HT, BioTek, Winooski, VT, USA) at 570 nm (OD570). To test blood compatibility, the possible hemolysis effect that was induced by nanoparticles was analyzed. Briefly, rat whole blood was collected, washed, and then diluted ten times with in volume with PBS first. 0.3 mL diluted red blood cell solution was mixed with 1.2 mL test sample solution and gently shaken before incubation at 37 ◦C for 2 h. Deionized water and PBS were used as the positive and negative controls, respectively. After centrifugation, the absorbance of the supernatant solution was recorded using a UV-Vis spectrophotometer from 500 to 650 nm. The difference in solution absorbance at 540 nm (OD540) between the sample and the negative control was divided by the difference in OD540 between the positive control and the negative control to calculate the hemolysis ratio.

3.8. In Vitro Blood Clot Lysis For static clot lysis, a blood clot was prepared from rat whole blood after diluting with equal volume PBS. The diluted blood (0.5 mL) was mixed with 100 µL thrombin solution (5 U/mL) in 1.8% NaCl and 0.3% CaCl2 solution in a 4 mL vial and pipetting for 20 times to form the clot. After 30 min, the vial was washed twice with PBS and 50 µL of sample (25 µU rtPA activity) was added to the vial Int. J. Mol. Sci. 2020, 21, 2690 20 of 23

containing 2 mL PBS. The vial was placed in a 37 ◦C oven for 10 min before measuring absorbance of the supernatant solution at 415 nm using a UV-Vis spectrophotometer. For the sample under magnetic guidance, a magnet (80 mm 50 mm 10 mm, 1700 gauss) was placed at the bottom of the vial during × × the incubation period. For dynamic clot lysis, 1 mL diluted blood was mixed with 0.2 mL thrombin solution and pipetting for 20 times. The blood solution was placed in a silicone tube (4 mm diameter) and then incubated at room temperature for 1 h. After clot formation, the tube was cut into 5-mm long pieces and the blood clot was flushed out from the tube with PBS. The blood clot was placed in a 4 mL vial and 50 µL of sample (50 µU rtPA activity) was added to the vial containing 4 mL PBS. The vial was incubated in a 37 ◦C or 43 ◦C oven in a Multi Bio RS-24 programmable rotator at 10 rpm for 10 min and the absorbance of the supernatant solution at 415 nm was measured using a UV-Vis spectrophotometer. The experiment was performed at 37 ◦C using a water bath for clot lysis in a flow model in a flow system employing flow occlusion induced by a blood clot (Figure 12a). A 4 mm diameter 5 mm × length blood clot prepared above was lodged tightly within a reduction point at the bottom of a tube where the inner diameter decreased from 4 mm to 3 mm. A fluid pressure gradient was generated to start flow of PBS using a syringe pump at 0.5 mL/min from the top while a magnet 80 mm 50 mm × × 10 mm, 1700 gauss) was placed below the clot for magnetic guidance. At time zero, different samples were introduced into the flow system by injecting 200 µL PBS, pPMNP, rtPA, or pPMNP-rtPA solution (100 µU rtPA) from a side opening and the reperfusion time for dissolution of the clot was measured to determine the lysis efficiency.

3.9. In Vivo Thrombolysis Using Rat Embolic Model A rat embolic model was used with the protocols that the Institutional Animal Care and Use Committee of Chang Gung University approved to determine the thrombolysis efficacy in vivo. Male SD rats were anesthetized by intraperitoneal injection with Inactin® (100 mg/kg). Mean arterial pressure was measured by carotid artery cannulation and a pressure transducer. The right iliac artery was cannulated with the tip of the catheter and reached to the bifurcation of abdominal aorta. A 3 mm 3 mm blood clot was introduced from the right iliac artery, and lodged in the left iliac artery upstream × of the pubic epigastric artery. The sample (pPMNP, rtPA, and pPMNP-rtPA) was administered via the right iliac artery and guided by a 0.5 T magnet from the abdominal aorta and iliac artery branch to the iliac artery and the femoral artery for 1 min, followed by magnetic guiding to the lower extremities for 4 min. The guiding cycle was repeated four times. The rate of iliac blood flow (IBF) and aortic blood flow (ABF) were determined by placing ultrasonic flow probes (T206, Transonic System Inc., Ithaca, NY, USA) on the left iliac artery and the aorta upstream of left renal artery, respectively. Real-time, high-resolution blood flow images at different time points after treatment were carried out by laser speckle contrast analysis while using a MoorFLPI laser speckle contrast imager from Moor Instruments (Devon, UK).

3.10. Statistical Analysis All of the results were presented as mean standard deviation (SD), except for in vivo experiments, ± which were presented as mean standard error of mean (SEM). For statistical analysis, the data are ± subject to one-way analysis of variance (ANOVA) and Tukey’s test with statistical significance declared with a p value of less than 0.05.

4. Conclusions In the study, we demonstrated the preparation of functional PLGA MNP modified through avidin-biotin interaction for the immobilization of fibrin-avid peptide (GPRPPGGSKGC) and thrombolytic drug (rtPA) in dual targeted thrombolytic therapy in vitro and in vivo. Iron oxide magnetic nanoparticles were well encapsulated in PLGA MNP, as confirmed by physico-chemical characterization, which allowed for drug delivery through magnetic guidance. Under the optimized Int. J. Mol. Sci. 2020, 21, 2690 21 of 23 preparation conditions, rtPA retains the same clot-lysis efficiency after conjugation to the dual targeted nanoparticle (pPMNP-rtPA) to facilitate its use in thrombolytic therapy. The magnetic guiding and fibrin binding ability associated with the dual targeted nanoparticle further reveal the preference of pPMNP-rtPA over free rtPA for in vitro and in vivo thrombolysis. The dual targeting thrombolytic nano-drug will have the potential to improve the efficacy of a thrombolytic drug and facilitate in vivo thrombolysis for future clinical application.

Author Contributions: H.-A.C., Y.-H.M., T.-Y.H. and J.-P.C. conceived and designed the experiments; H.-A.C. and T.-Y.H performed the experiments; H.-A.C., Y.-H.M., T.-Y.H. and J.-P.C. analyzed the data; H.-A.C. and J.-P.C. wrote the paper. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Ministry of Science and Technology, Taiwan grant number NERPD2C0461. Acknowledgments: We would like to express our appreciation for the financial support provided by Chang Gung Memorial Hospital (CMRPD250023, CMRPD1E0011 and BMRP249). The Microscopy Center in Chang Gung University is acknowledged for the transmission electron microcopy analysis. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Shiber, J.R.; Fontane, E.; Adewale, A. Stroke registry: Hemorrhagic vs ischemic strokes. Am. J. Emerg. Med. 2010, 28, 331–333. [CrossRef][PubMed] 2. Zivin, J.A. Acute stroke therapy with tissue plasminogen activator (tPA) since it was approved by the U.S. Food and Drug Administration (FDA). Ann. Neurol. 2009, 66, 6–10. [CrossRef][PubMed] 3. Gurman, P.; Miranda, O.R.; Nathan, A.; Washington, C.; Rosen, Y.; Elman, N.M. Recombinant tissue plasminogen activators (rtPA): A review. Clin. Pharmacol. Ther. 2015, 97, 274–285. [CrossRef] 4. Burrows, F.E.; Bray, N.; Denes, A.; Allan, S.M.; Schiessl, I. Delayed reperfusion deficits after experimental stroke account for increased pathophysiology. J. Cereb. Blood Flow Metabol. 2015, 35, 277–284. [CrossRef] [PubMed] 5. Gravanis, I.; Tsirka, S.E. Tissue-type plasminogen activator as a therapeutic target in stroke. Expert Opin. Ther. Targets 2008, 12, 159–170. [CrossRef][PubMed] 6. Zamanlu, M.; Farhoudi, M.; Eskandani, M.; Mahmoudi, J.; Barar, J.; Rafi, M.; Omidi, Y. Recent advances in targeted delivery of tissue plasminogen activator for enhanced thrombolysis in ischaemic stroke. J. Drug Target 2018, 26, 95–109. [CrossRef] 7. Doane, T.L.; Burda, C. The unique role of nanoparticles in nanomedicine: Imaging, drug delivery and therapy. Chem. Soc. Rev. 2012, 41, 2885–2911. [CrossRef] 8. Bonnard, T.; Gauberti, M.; Martinez de Lizarrondo, S.; Campos, F.; Vivien, D. Recent advances in nanomedicine for ischemic and hemorrhagic stroke. Stroke 2019, 50, 1318–1324. [CrossRef] 9. Chu, M.; Shao, Y.; Peng, J.; Dai, X.; Li, H.; Wu, Q.; Shi, D. Near-infrared laser light mediated cancer therapy

by photothermal effect of Fe3O4 magnetic nanoparticles. Biomaterials 2013, 34, 4078–4088. [CrossRef] 10. Kim, J.Y.; Kim, J.K.; Park, J.S.; Byun, Y.; Kim, C.K. The use of PEGylated liposomes to prolong circulation lifetimes of tissue plasminogen activator. Biomaterials 2009, 30, 5751–5756. [CrossRef] 11. Yu, M.K.; Park, J.; Jon, S. Targeting strategies for multifunctional nanoparticles in cancer imaging and therapy. Theranostics 2012, 2, 3–44. [CrossRef][PubMed] 12. Ma, Y.-H.; Liu, C.-H.; Liang, Y.; Chen, J.-P.; Wu, T. Targeted delivery of plasminogen activators for thrombolytic therapy: An integrative evaluation. Molecules 2019, 24, 3407. [CrossRef][PubMed] 13. Ma, Y.-H.; Wu, S.-Y.; Wu, T.; Chang, Y.-J.; Hua, M.-Y.; Chen, J.-P. Magnetically targeted thrombolysis with recombinant tissue plasminogen activator bound to polyacrylic acid-coated nanoparticles. Biomaterials 2009, 30, 3343–3351. [CrossRef][PubMed] 14. Chen, J.-P.; Yang, P.-C.; Ma, Y.-H.; Wu, T. Characterization of chitosan magnetic nanoparticles for in situ delivery of tissue plasminogen activator. Carbohydr. Polym. 2011, 84, 364–372. [CrossRef] 15. Chen, J.-P.; Yang, P.-C.; Ma, Y.-H.; Tu, S.-J.; Lu, Y.-J. Targeted delivery of tissue plasminogen activator by binding to silica-coated magnetic nanoparticle. Int. J. Nanomed. 2012, 7, 5137–5149. [CrossRef] 16. Hsu, H.-L.; Chen, J.-P. Preparation of thermosensitive magnetic liposome encapsulated recombinant tissue plasminogen activator for targeted thrombolysis. J. Magn. Magn. Mater. 2017, 427, 188–194. [CrossRef] Int. J. Mol. Sci. 2020, 21, 2690 22 of 23

17. Liu, C.-H.; Hsu, H.-L.; Chen, J.-P.; Wu, T.; Ma, Y.-H. Thrombolysis induced by intravenous administration of plasminogen activator in magnetoliposomes: Dual targeting by magnetic and thermal manipulation. Nanomed. Nanotechnol. Biol. Med. 2019, 20, 101992. [CrossRef] 18. McCarthy, J.R.; Sazonova, I.Y.; Erdem, S.S.; Hara, T.; Thompson, B.D.; Patel, P.; Botnaru, I.; Lin, C.P.; Reed, G.L.; Weissleder, R.; et al. Multifunctional nanoagent for thrombus-targeted fibrinolytic therapy. Nanomed. Nanotechnol. Biol. Med. 2012, 7, 1017–1028. [CrossRef] 19. Yurko, Y.; Maximov, V.; Andreozzi, E.; Thompson, G.L.; Vertegel, A.A. Design of biomedical nanodevices for dissolution of blood clots. Mater. Sci. Eng. C 2009, 29, 737–741. [CrossRef] 20. Hua, X.; Zhou, L.; Liu, P.; He, Y.; Tan, K.; Chen, Q.; Gao, Y.; Gao, Y. In vivo thrombolysis with targeted microbubbles loading tissue plasminogen activator in a rabbit femoral artery thrombus model. J. Thromb. Thromb. 2014, 38, 57–64. [CrossRef]

21. Zhou, J.; Guo, D.; Zhang, Y.; Wu, W.; Ran, H.; Wang, Z. Construction and evaluation of Fe3O4-based plga nanoparticles carrying rtpa used in the detection of thrombosis and in targeted thrombolysis. ACS Appl. Mater. Interfaces 2014, 6, 5566–5576. [CrossRef][PubMed] 22. Absar, S.; Kwon, Y.M.; Ahsan, F. Bio-responsive delivery of tissue plasminogen activator for localized thrombolysis. J. Control. Release 2014, 177, 42–50. [CrossRef][PubMed] 23. Juenet, M.; Aid-Launais, R.; Li, B.; Berger, A.; Aerts, J.; Ollivier, V.; Nicoletti, A.; Letourneur, D.; Chauvierre, C. Thrombolytic therapy based on fucoidan-functionalized polymer nanoparticles targeting P-selectin. Biomaterials 2018, 156, 204–216. [CrossRef][PubMed] 24. Huang, T.; Li, N.; Gao, J. Recent strategies on targeted delivery of thrombolytics. Asian J. Pharm. Sci. 2019, 14, 233–247. [CrossRef][PubMed] 25. Aruva, M.R.; Daviau, J.; Sharma, S.S.; Thakur, M.L. Imaging thromboembolism with fibrin-avid 99mTc-peptide: Evaluation in swine. J. Nuclear Med. 2006, 47, 155–162. 26. Thakur, M.L.; Pallela, V.R.; Consigny, P.M.; Rao, P.S.; Vessileva-Belnikolovska, D.; Shi, R. Imaging vascular thrombosis with 99mTc-labeled fibrin alpha-chain peptide. J. Nuclear Med. 2000, 41, 161–168. 27. Zhao, Y.; Jiang, Y.; Lv, W.; Wang, Z.; Lv, L.; Wang, B.; Liu, X.; Liu, Y.; Hu, Q.; Sun, W.; et al. Dual targeted nanocarrier for brain ischemic stroke treatment. J. Control. Release 2016, 233, 64–71. [CrossRef] 28. Luo, Y.; Yang, H.; Zhou, Y.-F.; Hu, B. Dual and multi-targeted nanoparticles for site-specific brain drug delivery. J. Control. Release 2020, 317, 195–215. [CrossRef] 29. Zhang, W.; Yu, Z.-L.; Wu, M.; Ren, J.-G.; Xia, H.-F.; Sa, G.-L.; Zhu, J.-Y.; Pang, D.-W.; Zhao, Y.-F.; Chen, G. Magnetic and folate functionalization enables rapid isolation and enhanced tumor-targeting of cell-derived microvesicles. ACS Nano 2017, 11, 277–290. [CrossRef] 30. Cui, Y.; Zhang, M.; Zeng, F.; Jin, H.; Xu, Q.; Huang, Y. Dual-targeting magnetic plga nanoparticles for codelivery of paclitaxel and curcumin for brain tumor therapy. ACS Appl. Mater. Interfaces 2016, 8, 32159–32169. [CrossRef] 31. Saengruengrit, C.; Ritprajak, P.; Wanichwecharungruang, S.; Sharma, A.; Salvan, G.; Zahn, D.R.T.; Insin, N. The combined magnetic field and iron oxide-PLGA composite particles: Effective protein antigen delivery and immune stimulation in dendritic cells. J. Colloid Interface Sci. 2018, 520, 101–111. [CrossRef][PubMed] 32. Boyer, C.; Priyanto, P.; Davis, T.P.; Pissuwan, D.; Bulmus, V.; Kavallaris, M.; Teoh, W.Y.; Amal, R.; Carroll, M.; Woodward, R.; et al. Anti-fouling magnetic nanoparticles for siRNA delivery. J. Mater. Chem. 2010, 20, 255–265. [CrossRef] 33. Livnah, O.; Bayer, E.A.; Wilchek, M.; Sussman, J.L. Three-dimensional structures of avidin and the avidin-biotin complex. Proc. Natl. Acad. Sci. USA 1993, 90, 5076–5080. [CrossRef][PubMed] 34. Northrop, B.H.; Frayne, S.H.; Choudhary, U. Thiol-maleimide “click” chemistry: Evaluating the influence of solvent, initiator, and thiol on the reaction mechanism, kinetics, and selectivity. Polym. Chem. 2015, 6, 3415–3430. [CrossRef] 35. Ravasco, J.M.J.M.; Faustino, H.; Trindade, A.; Gois, P.M.P. Bioconjugation with Maleimides: A Useful Tool for Chemical Biology. Chem. Eur. J. 2019, 25, 43–59. [CrossRef] 36. Wu, S.-L.; Jiang, H.; Hancock, W.S.; Karger, B.L. Identification of the unpaired cysteine status and complete mapping of the 17 disulfides of recombinant tissue plasminogen activator using LC-MS with electron transfer dissociation/collision induced dissociation. Anal. Chem. 2010, 82, 5296–5303. [CrossRef] 37. Absar, S.; Nahar, K.; Kwon, Y.M.; Ahsan, F. Thrombus-targeted nanocarrier attenuates bleeding complications associated with conventional thrombolytic therapy. Pharm. Res. 2013, 30, 1663–1676. [CrossRef] Int. J. Mol. Sci. 2020, 21, 2690 23 of 23

38. Chen, J.-P.; Liu, C.-H.; Hsu, H.-L.; Wu, T.; Lu, Y.-J.; Ma, Y.-H. Magnetically controlled release of recombinant tissue plasminogen activator from chitosan nanocomposites for targeted thrombolysis. J. Mater. Chem. B 2016, 4, 2578–2590. [CrossRef] 39. Lin, H.; Xu, L.; Yu, S.; Hong, W.; Huang, M.; Xu, P. Therapeutics targeting the fibrinolytic system. Exp. Mol. Med. 2020, 1–13. [CrossRef] 40. Tang, Z.; Li, D.; Wang, X.; Gong, H.; Luan, Y.; Liu, Z.; Brash, J.L.; Chen, H. A t-PA/nanoparticle conjugate with fully retained enzymatic activity and prolonged circulation time. J. Mater. Chem. B 2015, 3, 977–982. [CrossRef] 41. Lalatonne, Y.; Richardi, J.; Pileni, M.P. Van der Waals versus dipolar forces controlling mesoscopic organizations of magnetic nanocrystals. Nat. Mater. 2004, 3, 121–125. [CrossRef] 42. Singh, M.; Ulbrich, P.; Prokopec, V.; Svoboda, P.; Šantavá, E.; Štˇepánek, F. Effect of hydrophobic coating on

the magnetic anisotropy and radiofrequency heating of γ-Fe2O3 nanoparticles. J. Magn. Magn. Mater. 2013, 339, 106–113. [CrossRef] 43. Danaei, M.; Dehghankhold, M.; Ataei, S.; Hasanzadeh Davarani, F.; Javanmard, R.; Dokhani, A.; Khorasani, S.; Mozafari, M.R. Impact of particle size and polydispersity index on the clinical applications of lipidic nanocarrier systems. Pharmaceutics 2018, 10, 57. [CrossRef][PubMed] 44. Hermanson, G.T. Chapter 11—(Strept)avidin-Biotin Systems. In Bioconjugate Techniques, 3rd ed.; Hermanson, G.T., Ed.; Academic Press: Boston, MA, USA, 2013; pp. 465–505. [CrossRef]

45. Caruntu, D.; Caruntu, G.; O’Connor, C.J. Magnetic properties of variable-sized Fe3O4 nanoparticles synthesized from non-aqueous homogeneous solutions of polyols. J. Phys. D Appl. Phys. 2007, 40, 5801–5809. [CrossRef] 46. Wahajuddin Arora, S. Superparamagnetic iron oxide nanoparticles: Magnetic nanoplatforms as drug carriers. Int. J. Nanomed. 2012, 7, 3445–3471. [CrossRef][PubMed] 47. Torno, M.D.; Kaminski, M.D.; Xie, Y.; Meyers, R.E.; Mertz, C.J.; Liu, X.; O’Brien, W.D., Jr.; Rosengart, A.J. Improvement of in vitro thrombolysis employing magnetically-guided microspheres. Thromb. Res. 2008, 121, 799–811. [CrossRef][PubMed] 48. Gentile, P.; Chiono, V.; Carmagnola, I.; Hatton, P.V. An overview of poly(lactic-co-glycolic) acid (PLGA)-based biomaterials for bone tissue engineering. Int. J. Mol. Sci. 2014, 15, 3640–3659. [CrossRef] 49. McCarthy, J.R.; Patel, P.; Botnaru, I.; Haghayeghi, P.; Weissleder, R.; Jaffer, F.A. Multimodal nanoagents for the detection of intravascular thrombi. Bioconjug Chem. 2009, 20, 1251–1255. [CrossRef] 50. Runge, M.S.; Quertermous, T.; Zavodny, P.J.; Love, T.W.; Bode, C.; Freitag, M.; Shaw, S.Y.; Huang, P.L.; Chou, C.C.; Mullins, D.; et al. A recombinant chimeric plasminogen activator with high affinity for fibrin has increased thrombolytic potency in vitro and in vivo. Proc. Natl. Acad. Sci. USA 1991, 88, 10337–10341. [CrossRef] 51. Chapurina, Y.E.; Drozdov, A.S.; Popov, I.; Vinogradov, V.V.; Dudanov, I.P.; Vinogradov, V.V. Streptokinase@alumina nanoparticles as a promising thrombolytic colloid with prolonged action. J. Mater. Chem. B 2016, 4, 5921–5928. [CrossRef] 52. Liu, S.; Feng, X.; Jin, R.; Li, G. Tissue plasminogen activator-based nanothrombolysis for ischemic stroke. Expert Opin. Drug Deliv. 2018, 15, 173–184. [CrossRef][PubMed] 53. Whiting, D.; Dinardo, J. TEG and ROTEM: Technology and clinical applications. Am. J. Hematol. 2014, 89. [CrossRef][PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).